Geochemistry of tholeiites from Lanai, Hawaii...bedded (0.5-1 m thick), highly vesicular pahoehoe flows. We also collected a sample from Poaiwa, located on the northeast coast of Lanai,

Contrib Mineral Petrol (1992) 112: 520-542 Contributions to Mineralogy and Petrology �9 Springer-Verlag 1992

Geochemistry of tholeiites from Lanai, Hawaii H.B. West 1, M.O. Garcia z, D.C. Gerlach 3., and J. Romano 2.*

1 Hawaii Institute of Geophysics, University of Hawaii, Honolulu, HI 96822, USA 2 Department of Geology and Geophysics, University of Hawaii, Honolulu, HI 96822, USA 3 Lawrence Livermore National Laboratory, Livermore, CA 94550, USA

Received April 1, 1991/Accepted May 18, 1992

Abstract. Lanai is the third smallest of the fifteen principal subaerial shield volcanoes of the Hawaiian hotspot. This volcano apparently became extinct during the shield- building stage of volcanism, as shown by the absence of both alkalic cap and post-erosional lavas. Major and trace element analyses of 22 new samples collected primarily from 3 stratigraphic sections show that Lanai tholeiites span a large range in composition. Some Lanai lavas are unique geochemically among Hawaiian tholeiites in having the lowest abundances of incompatible trace elements of any Hawaiian lavas and well-developed positive Eu anomalies. The geochemical characteristics of these low- a.bundance Lanai tholeiites are not the result of alteration, differences in mantle source modal mineralogy, the presence of residual accessory mantle phases or fractional crystallization of such phases, assimilation of depleted [MORB] wall-rock, or accumulation/resorption of phenocrysts or xenocrysts. Incompatible trace element ratios (e.g., Nb/La, Nb/Th, La/Th, La/Hf, Ce/Pb) in Lanai tholeiites span considerable ranges and form coherent trends with each other and with absolute abundances of these elements. Large variations in La/Sm, La/Yb, and absolute REE abundances at constant MgO suggest that Lanai tholeiites formed by variable amounts of partial melting. However, large ranges in incompatible element ratios cannot be explained solely by variations in partial melting of a geochemically homogeneous source, but must reflect geochemical heterogeneities in the Lanai source. Partial melting modeling indicates that the mixed Lanai source is probably LREE-enriched [i.e., (La/Yb)c N > 1]. One component in the Lanai source, exemplified by the low- abundance tholeiites, has markedly lower REE/ HFSE, Th/HFSE, alkali/HFSE, and Ce/Pb ratios than other Lanai or Hawaiian tholeiites and may indicate the presence of recycled residual subduction zone materials in the Hawaiian plume source. The positive Eu anomalies

* Now at Charles Evans and Associates, 301 Chesapeake Drive, Redwood City, CA 94063, USA ** Now at Masa Fujioka and Associates, 99-1205 Halawa Valley Street, Aiea, Hawaii 96701, USA Correspondence to: H.B. West

that characterize the low-abundance Lanai tholeiites are not the result of plagioclase accumulation or assimilation but are a feature of this source component. Progressive temporal geochemical variations in Lanai tholeiites from 2 stratigraphic sections indicate that the source composition of these lavas probably evolved over time. This change could have resulted from a progressive decrease in the extent of partial melting of the Lanai source. The compositional variability of Lanai tholeiites suggests that geochemical heterogeneities in their source are larger than the scale of partial melting. Lanai tholeiites could not have formed by smaller degrees of partial melting of plume material than did the larger-volume Hawaiian shields. Therefore, volume differences between Haw~tiian shields must be controlled primarily by differences in the volume of supplied plume material rather than by differences in the degree of partial melting. The premature cessation of eruptive activity at Lanai may be attributed to relatively large degrees of partial melting of a small plume.

Introduction

Geochemical studies of Hawaiian lavas have contributed significantly towards understanding the nature and evolution of the suboceanic mantle. New insights are gained into the cause and extent of isotopic and geochemical heterogeneity in the source of Hawaiian hotspot volcanism as more volcanoes are studied and others examined in more detail. Results of early reconnaissance studies of Hawaiian volcanoes (e.g., Tatsumoto 1978)suggested that isotopic ratios for each volcano were distinct and spanned restricted ranges. Recent studies have shown that the evolution of Hawaiian volcanoes (i.e., shield-building, alkalic cap/post-shield, post-erosional/rejuvenation) is accompanied by systematic isotopic variations (Chen and Frey 1983; West and Leeman 1987; Kurz and Kammer 1991). Recent investigations of Kahoolawe (West et al. 1987; Leeman et al. in press) and Mauna Loa (Kurz and Kammer 1991) show that, even within the shield-building stage, geochemical and isotopic variability can be significant.

521

Strontium, lead, and neodymium isotope data for shield-building lavas from Hawaii form linear arrays thought to result from mixing between isotopically distinct source components (Stile et al. 1983, 1986; Roden et al. 1984; Staudigel et al. 1984; Feigenson 1984, 1986; Hegner et al. 1986; West et al. 1987; Tatsumoto et al. 1987; Kurz and Kammer 1991). An important approach to delineating the nature of these source components is to examine lavas from volcanoes that define the end-members of these arrays. Results from a reconnaissance Pb, Sr, and Nd isotope study of shield-building tholeiites from the Maui Volcanic Complex showed that" Lanai exemplifies one extreme of the Hawaiian shield-building array (West et al. 1987). The exact nature of the source component(s) representing this end-member is uncertain; both primitive mantle/bulk silicate Earth (the "PM" component; see White 1985) and enriched mantle (e.g., the "EMI" component; Zindler and Hart 1986) components have been suggested as possibilities.

The Hawaiian hotspot has produced shield volcanoes with volumes ranging from 6.4 x 103 km 3 (West Maui) to

42.5 • 103 km 3 (Mauna Loa) (Bargar and Jackson 1974). This range is even more extreme if the leeward volcanoes of the Hawaiian Chain are included. Any explanation for the cause of variations in shield volumes for hotspot volcanoes has implications concerning the relationship between volcano size and partial melting in the hotspot source. For example, volume variations between shield edifices could be related to differences in the extent of partial melting of similar volumes of plume material or to differences in the volume of plume material supplying each volcano. It has, however, been difficult to evaluate these possibilities because the geochemistry of shield- building lavas from the small-volume Hawaiian volcanoes has not been studied as extensively as that of the larger Hawaiian shields. Lanai's small size and distinct isotopic character make this volcano an excellent candidate for assessing geochemical heterogeneity in the Hawaiian plume source.

To assess the nature of components in the source of Hawaiian lavas and the geochemical heteogeneity of small-volume Hawaiian shield volcanoes, we collected

157 ~ 00 ' W 156~ ' I I I I

[]Craters �9 Cones

. ~ Dikes

< - / O u t e r caldera boundry

Contour interval 50Oft I E i

0 5km

KALAMANL

158~W 157 ~ I { I I

I KAHOOL IAWE ~ I I

I i

DNA I ~hlA

20 ~ 55' N

20 ~ 45 '

Fig. 1. Map of the island of Lanai (after Stearns 1940) showing locations of the three sampled stratigraphic sections (Kalamanui Gulch, Wawaeku, and Manele Bay) and other sample locations (Poaiwa, Maunalei Gulch). Inset map shows the location of Lanai

relative to other volcanoes of the Maui Volcanic Complex: East and West Molokai, West Maui, Haleakala (East Maui), and Kahoolawe (after West et al. 1987). Solid line around the islands is the 180 m submarine contour

522

lava samples from three strat igraphic sections on Lanai and analyzed them for major and trace element compositions. The purpose of this paper is to (1) show that some Lanai tholeiites are geochemically distinct from other Hawai ian tholeiites, (2) examine possible mel t ing models for the origin of these lavas, and (3) evaluate the na ture of geochemical heterogeneities in the source of Hawai ian shield-building lavas.

Brief geologic description, previous investigations, and sampling

The island of Lanai, which rises to a maximum elevation of 1027 m above sea level, is one of 6 shield volcanoes that form the Maul Volcanic Complex (Fig. 1). Lanai is the third smallest of the fifteen principal Hawaiian subaerial shield volcanoes, with an estimated volume of 12 • 103 km 3 (Bargar and Jackson 1974). This volcano has three rift zones, trending south, southwest, and north-west, that exend outwards from the central Palawai caldera (Fig. 1). The principal period of shield-building volcanism culminated in collapse of the caldera, followed by sporadic volcanic activity on the southern portion of the island and along the northwest rift zone (Stearns 1940). Lanai appears to be the only extinct volcano within the principal Hawaiian Islands without surface exposures of alkalic cap or post-erosional lavas.

Potassium-argon ages for six lava samples collected from the flanks of Lanai range from 1.20 • 0.17 Ma to 1.46 4- 0.25 Ma (Bonhommet et al. 1977), although younger ages with large relative errors were reported for 3 samples collected from Maunalei Gulch in the island's interior (0.71 • 1.27 to 0.86 • 0.55 Ma; Naughton et al. 1980). Five major element analyses of Lanai lavas have been published previously (Washington 1923; Bonhommet et al. 1977), including 4 of the age-dated samples (Bonhommet et al. 1977). These four samples were analyzed for trace elements by instrumental neutron activation analysis (INAA) (Budahn and Schmitt 1985) and for Sr, Pb, and Nd isotopes (West et al. 1987). Results of the isotope study revealed that tholeiites from Lanai, along with those from Koolau and Kahoolawe, define an end-member for Hawaiian lavas. Lavas from these volcanoes have less radiogenic Pb and more radiogenic Sr and Nd than lavas from other Hawaiian volcanoes (Fig. 2).

We collected twenty lava samples from three stratigraphic sections on Lanai: the south coast (Manele Bay; ~ 163 m of section), the east flank (Wawaeku Gulch; ~ 43 m of section), and the west coast (Kalamanui Gulch ~ 20 m of section) (Fig. 1). Lavas exposed in these sections consist of massive (4-8 m thick) a'a flows and thin- bedded (0.5-1 m thick), highly vesicular pahoehoe flows. We also collected a sample from Poaiwa, located on the northeast coast of Lanai, and obtained a split of a sample (reported in Naughton et al. 1980) from Maunalei Gulch, the deepest section exposed on Lanai.

Analytical methods

Samples were analyzed for major and trace (Nb, Zr, Y, Ba, Rb, Sr, Sc, V, Cr, Ni, Zn, Cu) element compositions by X-ray fluorescence (XRF) at the University of Hawaii following the method of Norrish and Chappell (1977). The University of Hawaii XRF facility consists of a fully-automated, wavelength-dispersive Siemens SRS 303 AS. The major element analyses have about a 1% relative error, based on repeated analyses of standard rocks.

Abundances ofTi, Mn, Cr, Ni, Sc, V, Co, Cu, Zn, Ga, Mo, Sn, Cs, Rb, Sr, Ba, REE, Y, Zr, Hf, Nb, Ta, U, Th, and Pb were determined by inductively coupled plasma emission mass spectrometry (ICP- MS) in the Nuclear Chemistry Division at Lawrence Livermore National Laboratory (LLNL) using analytical techniques developed there. Analyses were done on 2 instruments, a VG Elemental PlasmaQuad PQI and a VG Elemental PlasmaQuad PQ2. Detec-

tion limits are in the pg-g-~ range. The isotope masses used to determine elemental concentrations and estimated precision and accuracy, based on replicate analyses of BHVO-1, are listed in Table 1.

Results

Petrography

Samples from this study are similar petrographically to tholeiites from other Hawaiian volcanoes (e.g., Kilauea, Mauna Loa). Lanai tholeiites range from nearly aphyric ( < 0.1 vol.% total phenocrysts) to picritic ( > 15 vol.% olivine phenocrysts). The major phenocryst phase is olivine with minor or no plagioclase (0.0 2.4 vol.%) and/or clinopyroxene (0.~0.7 vol.%) (Table 2). In hand specimen, all samples appear unaltered. In thin-section, some samples exhibit minor ground-mass oxidation and contain olivines with thin ( < 0.01 mm) iddingsite rims. None of the samples contain secondary phases associated with alteration of basaltic lavas (e.g., zeolite, clay minerals, calcite).

Major and trace elements

The new samples, like the five samples analyzed by Washington (1923) and Bonhommet et al. (1977), are all tholeiites. Major element compositions of Lanai lavas (Tables 3A-3C) are similar to those of other Hawaiian tholeiites, except that most contain lower total

H a w a i i a n I s o t o p e S y s L e m a t i e s

0.7050

0.7045

0.7040

0.7035

0.7050

0.7025

\ EM1 PM "k

z,,~.r ~ ~ E M

I~MVC SB L~va~

J Eratee~la /~ . 51 Kilauea �9 I PSB L a v a s _ ~

HIMU t DM

L . . . . . . / . . . . . . . . . . . . . . . . 17.0 18.0 19.0 2~176

Fig. 2. Hawaiian isotope systematics for 2~176 versus STSr/S6Sr. Shown are fields for shield-building (SB) lavas from Lanai, other volcanoes of the Maui Volcanic Complex (MVC), and Kilauea, and for post shield-building (PSB) lavas from Haleakala. Hawaiian shield-building lavas form a linear array, with Lanai and Kilauea defining the end-members. The Kilauea end-member could represent an enriched mantle component close in composition to Kilauea lavas (e.g., "EM"; West and Leeman 1987) or one containing more radiogenic Pb (e.g,, "HIMU"; Zindler and Hart 1986). The Lanai end-member could represent a primitive mantle component ("PM"; White 1985) or one that is isotopically more extreme (e.g., "EM l"; Zindler and Hart 1986). Post shield-building lavas define an opposing trend, probably reflecting contamination of plume magmas by a depleted mantle ("DM") component (e.g., oceanic lithosphere or asthenosphere; Chen and Frey 1985; West and Leeman 1987). Data from: Tatsumoto (1978), Chert and Frey (1985), Hegner et al. (1986), Newsom et al. (1986), Stille et al. (1986), West and Leeman (1987), West et al. (1987), and Chen et al. (1991)

523

Table 1. Consensus values for BHVO-1 and the mean and percent standard deviation of three replicate analyses obtained by ICP-MS for this study. Also shown are the isotope masses used to determine elemental concentrations for ICP-MS. Data given in ppm

Consensus a Value +_ n ICP-MS % SD Masses b

Se 31.8 1.3 36 34.4 4.9 Ti 16220 380 31 16948 4.0 V 317 12 26 336 3.9 Cr 289 22 36 275 3.4 Mn 1300 62 43 1355 3.2 Ni 121 2 29 109 8.0 Cu 136 6 15 139 2.7 Zn 105 5 15 113 0.6 Ga 21 2 6 20.7 2.6 Rb 11 2 27 11.0 12.4 Sr 403 25 32 428 6.7 Y 27.6 1.7 22 27.1 15.6 Zr 179 21 27 176 5.2 Nb 19 2 19 19.6 0.5 Mo 1.02 0.10 9 1.16 1.3 Sn 2.1 0.5 8 2.1 29.8 Cs 0.13 0.06 8 0.102 8.8 Ba 139 14 37 139 10.9 La 15.8 1.3 53 16.0 4.5 Ce 39 4 56 42.8 4.9 Pr 5.7 0.4 9 6.17 7.0 Nd 25.2 2.0 45 25.7 6.0 Sm 6.2 0.3 53 6.23 4.5 Eu 2.06 0.08 50 2.18 4.3 Gd 6.4 0.5 31 6.63 2.5 Tb 0.96 0.08 35 0.880 4.7 Dy 5.2 0.3 28 4.94 2.3 Ho 0.99 0.08 16 0.888 5.4 Er 2.4 0.2 18 2.35 4.2 Tm 0.330 0.040 16 0.291 6.6 Yb 2.02 0.20 57 1.87 5.3 Lu 0.291 0.026 32 0.249 5.3 Hf 4.38 0.22 30 4.09 5.6 Ta 1.23 0.13 26 1.21 3.9 Pb 2.6 0.9 7 1.09 5.5

0.98 ~ Th 1.08 0.15 32 1.09 8.3 U 0.42 0.06 15 0.392 3.0

45 47, 49 51 52 55 60 63 66 69 85 88 89 90, 91 93 95,97

117, 118, 119, 120 133 137,138 139 140 141 143, 145 147, 149 151 157 159 161, 162, 163 165 166, 167 169 173, 174 175 178, 180 181 206, 207, 208

232 238

n, # of samples. a Gladney and Roelandts (1988), Govindaraju (1989) b Where more than one mass is listed, averages were used to establish concentration

Isotope dilution value, determined at the University of Hawaii (K. Spencer, analyst)

alkalies (Na20 + K20) at a given SiO 2 content. The A1203/CaO ratios of Lanai lavas are similar to values for Mauna Loa lavas, but they are distinctly higher ( > 1.3) than those of Kilauea tholeiites ( < 1.3) and lower than ratios in Koolau tholeiites (1.4-1.7). With decreasing MgO content, SiO2, A1203, CaO, Na20, and TiO 2 contents increase systematically. Total iron, as FeO, decreases with decreasing MgO (although the data are somewhat scattered), down to about 7 wt.% MgO, at which point FeO begins to increase. The K 2 0 and P20 5 contents and AIzO3/CaO ratios are not correlated significantly with MgO content.

Lanai tholeiites span a considerable range in incompatible trace element contents. For these lavas, abundances of these elements are not correlated with MgO content. Several samples contain ex- tremely low abundances of immobile incompatible trace elements (i.e., elements not affected significantly by surface alteration; e.g., Th < 0.5 ppm, Nb < 7 ppm, Zr < 100 ppm, La < 3 ppm, Sm < 2 ppm, Yb < 1 ppm; Table 3). These low-abundance tholeiites have the lowest REE concentrations of any Hawaiian lavas. They also have among the least-fractionated relative REE contents of any

Hawaiian lavas; e.g. (La/Sm)c N = 0.9-1.1 and (La/Yb)c N = 1.9~.8 (Figs. 3-4; CN = element abundances normalized to mean CI- chondrite; Anders and Grevesse 1989). Mauna Loa lavas also have low (La/Sm)cN (0.92-1.51), but they have slightly higher (La/Yb)cN ratios (2.5 3.9) than do the low-abundance Lanai tholeiites (based on Manna Loa data from BVSP 1981; Newsom et al. 1986; Tilling et al. 1987; Rhodes 1988). The low-abundance Lanai thoteiites have Th, light rare-earth element (LREE), Pb, Zr, and Hf contents that fall within the range for average normal mid-ocean ridge basalts (NMORB); heavy rare-earth element (HREE) and Y abundances are lower in these lavas than in NMORB (Fig. 5). The Cs, Rb, and U contents of the low-abundance Lanai tholeiites are also low; however, in some Lanai samples, abundances of these elements may have been affected by minor alteration (see the discussion in the following section).

The REE contents of the low-abundance Lanai lavas are not correlated with MgO (Figs. 3, 4a) and are significantly lower than REE abundances in other Hawaiian tholeiites of equivalent MgO content. For example, sample LMan-2 (7.2wt.% MgO, < 3%

524

Table 2. Modal mineralogy of tholeiites from Lanai, based on 1000 point counts for each sample

Olivine Clinopyroxene Plagioclase mph ph mph ph mph ph gm

Wawaeku LWaw-1 0.5 0.1 0.7 0.0 0.1 0.0 98.7 LWaw-2 0.9 0.5 0.4 0.3 0.9 0.6 96.4 LWaw-4 7.6 6.9 8.0 0.0 0.0 0.0 77.5 LWaw-4b 7.4 16.0 0.7 0.0 1.0 0.2 74.7 LWaw-5 2.7 4.4 1.0 0.0 12.7 0.0 79.2 LWaw-6 4.3 8.2 0.9 0.0 5.6 2.3 78.7 LWaw-7 3.5 11.4 1.5 0.0 8.1 0.6 74.9 LWaw-8 1.8 7.3 0.0 0.0 0.0 0.0 90.9

Manele LMan-2 0.9 0.9 0.2 0.1 1.4 1.7 94.9 LMan-3 1.0 0.5 0.3 0,2 4.1 1.7 92.2 LMan-4 3.4 4.1 0.6 0.0 0.3 0.7 90.9 LMan-5 0.8 0.5 0.5 0.4 0.0 0.0 97.8 LMan-6 0.3 0.3 1.9 0.0 . 5.8 0.1 91.6

Kalamanui LKal-1 2.6 5.8 0.0 0.0 0.1 0.0 91.5 LKal-3 2.3 1.9 0.1 0.2 0.0 0.0 95.5 LKal-4 3.6 5.3 0.0 0.0 0.0 0.0 91.1 LKal-5 4.1 0.1 0.0 0.0 0.0 0.0 95.8 LKal-6 0.6 0.1 1.4 0.7 8.1 2.4 86.7

Poaiwa LPoa-1 0.2 0.0 0.1 0.0 0.1 0.0 99.8

Maunalei KA-25 0.0 0.7 0.0 0.4 0.0 1.9 97.0

ph = phenocryst, mph = microphenocryst, gm = groundmass

IO0

.4~..m

L~

"0

0 ~I0 0

C.) 0 rc

I O0

. m

r O i - - 1 0 O

O or"

1

;g t - o r

O ,,4 O O

n -

, , , , , , , , 4 , , , , ,

Wawaeku!

kWaw-5 (8.0 wt.% MgO) [] LWaw-2 (6.5 wt.% MgO) �9 LWaw-6 (8.9 wt.% MgO) o LWaw-1 (6.1 wt.%MgO) A A LWaw-4b (14.8 wt.% MgO)

r i i i i i i i i r i r i

La Ce Pr NdPmSmEu Gd Tb Dy Ho Er TmYb Lu

i , i i , , 4 i , , i i ~ , ,

Manele

�9 CMan-6 (7.3 wt.% MgO) [] LMan-5 (7.6 wt.% MgO) �9 LMan-3 (7.,3 wt.% MgO) c LMan-4 (10.4 wt.% MgO) �9 LMan-2 (7.2 wt.% MgO) B

_ i i i i i i i 5 i i i p i

La Ce Pr NdPmSm Eu GdTb Ely H0 Er TmYb Lu

10

;5 o = c - 1 0

O

O n -

1 1

Kalamanui

- " A " - - . x , t ~ A ~ ' ~

�9 LKal-2 (6.2 wt.% MgO) ~ J ~ - ~ [] LKal-5 (6.7 wt.% MgO) �9 LKal-6 (6.7 wt.% MgO) o LKaI-8 (9.2 w1.% MgO) C * LKal-4 (10.9 wt.% MgO)

i i i i i i i J I i i i

La Ce Pr NdPmSmEu Gd Tb Dy Ho Er "I'm Yb Lu

, , , , i i i f , , i , , i f

Poaiwa & Maunalei

o LPoa-I (7.1 wt.% MgO) KA-25 (6.6 wt.% MgO) D

La Ce Pr Nd PmSm Eu Gd Tb Dy Ho Er TmYb Lu

Fig. 3A-D. Rare-earth element abundances in representative Lanai tholeiites normalized to mean Cl-chondrite (Anders and Grevesse 1989). A Wawaeku Gulch section. B Manele section. C Kalamanui Gulch section. D Maunalei Gulch and Poaiwa

Table 3A. Major and trace element data for tholeiites from element data are given in ppm

525

Wawaeku Gulch, Lanai. Major element data are given in weight percent. Trace

LWaw-1 LWaw-2 LWaw-4 LWaw-4b LWaw-5 LWaw-6 LWaw-7 LWaw-8

SiO 2 48.44 51.18 48.17 49.24 50.45 48.52 49.24 50.41 A1203 15.05 14.20 11.58 11.75 13.78 12.89 12.80 12.27 Fe20 3 12.40 13.00 12.53 12.34 12.41 12.91 12.82 12.67 MgO 5.86 6.48 14.48 14.79 7.93 11.82 11.69 12.48 CaO 9.41 9.87 8.04 8.56 10.06 8.82 8.91 8.84 Na20 1.85 2.01 1.17 1.19 1.63 1.48 1.46 1.23 K20 0.08 0.18 0.04 0.12 0.16 0.06 0.10 0.29 P20~ 0.15 0.21 0.12 0.13 0.20 0.15 0.17 0.19 TiO2 2.21 2.40 1.67 1.70 2.05 1.99 1.96 1.88 MnO 0.16 0.17 0.17 0.17 0.18 0.17 0.17 0.17 LOI 2.59 0.52 1.88 0.49 1.19 1.47 1.00 0.08 Total 98.13 100.21 99.81 100.48 100.01 100.25 100.31 100.51

Cs 0.021 0.031 0.004 0.019 0.007 0.004 0.001 0.060 Rb 0.18 1.27 0.11 1.54 0.38 0.10 0.15 5.07 Sr 374 380 243 259 338 293 286 332 Ba 45.9 78.0 32.8 34.6 83.1 52.6 61.5 108.0 La 4.97 8.41 4.52 4.67 10.07 7.68 7.27 9.46 Ce 12.56 21.38 10.88 12.42 23.72 14.61 16.07 28.18 Pr 2.12 3.75 1.91 2.00 4.10 3.44 3.30 4.52 Nd 9.40 16.00 8.30 9.00 17.10 14.60 14.50 21.00 Sm 2.65 4.33 2.34 2.55 4.60 4.10 4.06 6.85 Eu 1.22 1.70 0.99 1.06 1.70 1.78 1.66 2. 82 Gd 3.00 4.75 2.86 2.98 5.15 4.28 4.45 9.45 Tb 0.439 0.688 0.404 0.437 0.720 0,675 0.652 1.500 Dy 2.79 4.09 2.50 2.59 4.33 4,11 3.91 9.22 Ho 0.506 0.721 0.444 0.459 0.789 0,737 0.720 1.900 Er 1.42 1.99 1.30 1.25 2.15 1,98 1.96 4.90 Tm 0.188 0.269 0.172 0.180 0.274 0.279 0.261 0.662 Yb 1.20 1.71 1.05 1.09 1.74 1,71 1.72 4.12 Lu 0.169 0.227 0.152 0.144 0.239 0.230 0.221 0.587 Y 20.3 28.5 20.4 19.2 32.9 25.7 27.1 85.9 Y* 17.3 24.6 18.3 16.0 - 22.7 23.9 64.2 Hf 3.16 3.43 2.05 2.12 2.81 2.60 2.48 2.70 Zr 137.8 148.4 89.0 92.7 122 110 107.3 118.4 Nb 8.56 9.02 6.70 6.73 7.58 7.65 7.51 8.01 Pb 0.80 1.02 0.60 0.61 0.91 0.75 0.76 0.85 Th 0.436 0.504 0.337 0.352 0.441 0.423 0.415 0.440 U 0.063 0.123 0.060 0.073 0.130 0.079 0.094 0.164 Mo 0.42 0.78 0.37 0.49 0.69 0.41 0.52 0.86 Sn 3.31 2.62 1.91 2.78 2.65 2.95 1.75 2.55 Sc 37.7 37.3 33.3 31.9 37.2 31.8 34.0 32.2 V 310 353 256 264 308 263 267 258 Cr 305 226 669 662 398 685 716 573 Ni 95.9 95.1 715 682 159 439 444 429 Cu 91.4 92.1 86.5 103 116 96.2 103 77.8 Zn 111 114 94.1 101 103 99.8 100 120 Ga 17.2 17.4 12.3 12.5 15.9 13.8 13.9 14.7 Ti 15109 15993 10638 10739 12998 11483 11705 10973 Mn 1223 1308 1213 1287 1300 1230 1244 1284

phenocrysts) contains 2.6 ppm La, whereas tholeiites from other Hawaiian volcanoes with similar MgO contents have La abundances typically four or more times higher (Fig. 4). Note that picritic tholeiites from Kilauea with > 19 wt.% MgO have significantly higher incompatible element contents than do the low-abundance Lanai tholeiites (e.g., 6.9-11.5 ppm La; BVSP 1981; Casadevall and Dzurisin 1987a, 1987b; Tilling et al. 1987) (Fig. 4). Even 2 of the most incompatible element-poor Hawaiian tholeiites, a picrite from Mauna Loa with 23.7 wt.% MgO (Tilling et al. 1987) and another from Kahoolawe with 16.7 wt.% MgO (Leeman et al. in press), have higher La abundances (5.7 and 5.5 pprn La, respectively). Chen and Frey (1983) suggested that the most primitive Haleakala tholeiite in their study (sample C-122; 16.3 wt.% MgO) could represent a primary magma, yet that sample contains nearly three times as much La as some of the low-abundance Lanai tholeiites.

The low-abundance Lanai tholeiites have positive Eu anomalies (Fig. 3), a rare feature in Hawaiian rocks (Eu/Eu* = 1.03-1.61, where Eu* is the extrapolated Eu abundance taken between Sm and Gd), In Lanai lavas, Eu/Eu* is correlated with highly incompatible element abundances and several incompatible trace element ratios (e,g., La, Th, La/Yb, La/Sm, Nb/La, Nb/Th~ La/Th, Sr/Nd), but it is not well-correlated with most major elements or with elements and elemental ratios indicative of plagioclase control (e.g., SiO2, AlzO3, M g ~ , A1203/CaO, Al2Oa/TiO2, Sr). Sr is correlated strongly with Zr abundance [-correlation coefficient (r) = 0.955], indicating these elements were simiarly incompatible in the genesis of Lanai lavas.

Several highly incompatible trace element ratios in Lanai lavas span significantly larger ranges than have been recognized previously for Hawaiian tholeiites (e.g., Nb/Th 16-34, La/Th 9 23, Nb/La 0.75 3.11, Nb/U 49-103, Ce/Pb 12-33; excludino samples

526

Table 3B. Major and trace element data for tholeiites from Manele, Lanai. Major element data element data are given in ppm

are given in weight percent. Trace

LMan-2 LMan-3 LMan-4 LMan-5 LMan-6

SiO 2 52.14 53.00 51.11 52.96 52.35 AI20 3 13.80 14.05 12.92 13.19 13.87 Fe20a 11.70 11.46 12.11 11.82 11.58 MgO 7.17 7.29 10.45 7.58 7,22 CaO 10.01 9.86 9.53 9.65 9.88 Na20 1.79 1.77 1.58 1.67 1.73 K20 0.18 0.24 0.20 0.48 0.48 P205 0.09 0.13 0.I0 0.17 0.23 TiO 2 2.14 1.95 1.78 2.03 2.02 MnO 0.16 0.16 0.17 0.17 0.16 LOI 0.85 0.01 0.69 0.57 0.37 Total 100.03 99.91 100.63 100.28 99.89

Cs 0.032 0.038 0.029 0.067 0.071 Rb 2.71 3.74 2.76 6.74 6.24 Sr 330 334 266 331 335 Ba 63.7 67.4 52.4 85.9 81.0 La 2.63 4.23 3.60 6.92 9.12 Ce 7.33 11.51 9.63 18.43 23.81 Pr 1.21 1.86 1.57 2.95 3.77 Nd 5.50 8.40 6.90 12.30 16.20 Sm 1.75 2.49 2.08 3.49 4.55 Eu 1.03 1.22 1.04 1.26 1.59 Gd 2.14 2.87 2.51 3.81 4.79 Tb 0.344 0.434 0.391 0.543 0.707 Dy 2.04 2.70 2.41 3.33 4.19 Ho 0.377 0.487 0.432 0.579 0.740 Er 1.04 1.34 1.24 1.68 2.05 Tm 0.150 0.192 0.176 0.223 0.264 Yb 0.94 1.19 1.09 1.39 1.74 Lu 0.128 0.166 0.153 0.195 0.230 Y 14.2 19.1 17.1 22.8 29.2 Y* 11.8 16.1 14.8 19.5 24.7 Hf 2.93 2.65 2.12 2.80 2.81 Zr 124.0 116.2 88.7 118 116 Nb 8.18 7.72 6.20 8.57 8.41 Pb 0.61 0.69 0.59 0.95 0.89 Th 0.241 0.277 0.247 0.449 0.527 U 0.080 0.100 0.090 0.155 0.170 Mo 0.87 0.85 0.74 0.81 1.09 Sn 0.80 2.14 0.78 2.19 2.46 Sc 35.2 31.6 32.3 34.4 33.2 V 295 284 274 279 285 Cr 348 324 59 l 324 290 Ni 140 153 347 119 133 Cu 69.5 69.4 108 104 114 Zn 102 101 95.1 97.4 101 Ga 15.6 15.6 13.6 14.9 15.5 Ti 13015 12074 10563 12001 12063 Mn 1257 1265 1294 1262 1266

with low K20/P2Os, high Ba/Rb, and high K/Rb, which may have undergone minor alteration). In addition, some of these ratios in Lanai tholeiites vary systematically with absolute abundances of these elements (Fig. 6). These results refute the contention that ocean island basalts (OIB) and MORB possess constant Ce/Pb (25 _+ 5) and Nb/U (47 _+ I0) ratios that are independent of trace element abundances (Hofmann et al. 1986; Hofmann 1986, 1988; Newsom et al. 1986). This result is not unique to Lanai; tholeiites from Kahool- awe also display relatively large variations in Nb/Th ratios (9.6-32.5) that vary systematically with incompatible element abundances (Leeman et al. in press).

The low Ce/Pb ratios of some Lanai tholeiites approach estimated primitive mantle values (9.15-9.59; Hofmann 1988; Sun and

McDonough 1989). The samples with low-Ce/Pb have low La/Th ratios (as low as about 9) that are near-chondritic (chondritic La/Th approx. 8; Anders and Grevesse 1989) and La/Hf ratios that overlap chondritic values (chondritic La/Hf ~ 2.3; Anders and Grevesse 1989). The Ce/Pb ratios of unaltered samples are correlated strongly with Nb/La, Nb/Th, La/Sm, La/Yb, and La/Hf ratios. In contrast to the highly variable nature of some incompatible element ratios in these lavas, Zr/Hf ratios are relatively constant (43.2 • 1.4) and non-chondritic (chondritic Zr/Hf approx. 38; An- ders and Grevesse 1989). The Ba/Rb ratios of unaltered Lanai lavas (12-24) generally are higher than either the average OIB value of 12.45_ 0.30 (Hofmann and White 1983) or the primitive mantle value of 11.3 _+ 0.2 (Hofmann 1988).

527

Table 3C. Major and trace element data for tholeiites from Kalamanui gulch, Maunalei gulch, and Poaiwa. Major element data are given in weight percent. Trace element data are given in ppm

LKaI-1 LKal-2 LKal-3 LKal-4 LKal-5 LKal-6 LPoa-1 KA-25

SiO2 50.46 51.42 50.76 51.92 51.83 53.05 A120 3 12.98 13.56 12.33 14.49 13.74 13.64 Fe20 3 12.40 - 12.23 12.08 11.88 12.31 12.04 MgO 10.09 - 9.12 10.81 - 6.69 7.09 6.64 CaO 9.46 9.26 8.97 10.17 10.35 9.92 Na20 1.41 1.57 2.34 2.02 2.27 1.83 K20 0.07 - 0.15 0.18 - 0.25 0.35 0.39 P2Os 0.18 - 0.15 0.09 - 0.16 0.16 0.13 TiO2 2.12 1.66 1.66 2.05 2.18 2.12 MnO 0.17 0.17 0.14 0.17 0.13 0.17 LOI 1.06 - 1.09 0.20 - 0.53 0.01 0.39 Total 100.38 - 100.37 99.54 - 100.32 100.41 100.33

Cs 0.006 0.070 0.029 0.018 0.070 0.031 0.046 0.040 Rb 0.40 6.55 1.79 1.50 5.46 4.00 4.83 5.40 Sr 286 344 256 257 306 322 339 301 Ba 47.0 108.2 41.0 48.6 95.6 64.8 73.0 67.4 La 6.70 9.01 5.08 2.59 8.27 6.31 7.54 4.99 Ce 16.54 24.16 13.28 6.94 22.43 16.57 18.44 13.55 Pr 2.59 3.85 2.07 1.12 3.66 2.57 2.99 2.15 Nd 10.90 16.70 9.50 4.90 15.20 11.00 12.90 9.60 Sm 2.87 4.63 2.79 1.58 4.02 3.18 3.48 2.83 Eu 1.25 1.73 1.13 0.74 1.49 1.33 1.49 1.49 Gd 3.25 4.87 3.17 1.76 4.23 3.57 4.12 3.20 Tb 0.483 0.735 0.473 0.292 0.655 0.516 0.583 0.475 Dy 2.83 4.27 2.89 1.77 3.74 3.09 3.54 3.08 Ho 0.510 0.817 0.531 0.327 0.716 0.584 0.647 0.542 Er 1.34 2.11 1.44 0.94 1.84 1.60 1.79 1.60 Tm 0.198 0.291 0.203 0.132 0.239 0.227 0.248 0.219 Yb 1.16 1.84 1.24 0.82 1.60 1.38 1.54 1.43 Lu 0.165 0.259 0.178 0.116 0.226 0.194 0.216 0.191 Y 21.2 31.3 20.7 12.6 26.4 22.2 26.5 21.7 ya 18.1 18.1 11.2 - 19.6 23.0 19.1 Hf 2.47 2.91 2.08 1.99 2.72 2.62 2.79 2.79 Zr 115 119.0 87.6 88.4 I l 0.1 115.9 128.2 124.0 Nb 10.40 9.37 5.88 6.53 8.01 8.47 8.41 7.90 Pb 0.75 0.79 0.56 0.56 0.78 0.75 0.75 0.74 Th 0.392 0.474 0.337 0.289 0.444 0.451 0.441 0.365 U 0.105 0.152 0.084 0.097 0.130 0.127 0.138 0.150 Mo 0.70 1.28 0.65 0.56 0.91 0.79 0.75 1.02 Sn 3.12 2.36 194 1.59 1.75 1.77 2.38 1.21 Sc 32.7 30.9 29.2 31.1 31.8 31.8 V 302 - 247 227 288 299 290 Cr 553 - 395 547 - 257 260 265 Ni 302 98.6 176 286 153 135 110 92.4 Cu 105 116 118 111 49.8 105 106 114 Zn 114 118 102 93.2 107 101 108 103 Ga 14.4 16.4 13.3 12.7 14.6 15.2 15.8 15.5 Ti 12869 9354 9246 11704 12780 12145 Mn 1339 - 1313 1227 - 1226 1275 1235

Major elements determined by XRF at the University of Hawaii. Trace elements determined by ICP-MS at Lawrence Livermore National Laboratory. a Determined by XRF at the University of Hawaii

Alteration effects

Some Lanai samples have low alkali (Na20, K20, Rb, Cs) contents and low K20/P2Os ( < 0.9), high K20/Rb ( > 0.1), and high Ba/Rb ( > 25) ratios, compared to other samples (K20/P2Os: 0.9-3.0, KzO/Rb: < 0.1, Ba/Rb: 12-24) (Fig. 7a). Similar geochemical features in tholeiites from Haleakala and Kohala have been ascribed to mobilization of alkalies during subaerial alteration (Feigenson et al. 1983; Chen and Frey 1985). Almost all of the altered Lanai samples are from the Wawaeku section, located on the windward side of the island where rainfall is somewhat higher than at the other sampled

locations (samples LWaw-1, LWaw-2, LWaw-4, LWaw-5, LWaw-6, LWaw-7; sample LKaI-1, from the Kalamanui section, may also have been affected by minor subaerial alteration). Lavas from the wet, windward flank of Mauna Kea also were found to have lost alkalies relative to lavas from the dry, leeward flank of the volcano (Frey et al. 1990a), even though they are much younger than lavas from Lanai (0.1 0.3 Ma versus 0.7 1.5 Ma).

The negative correlation between K 2 0 / P 2 0 s and Ba/Rb ratios at K20/P205 < 1.0 probably results from alkali loss during alteration (Fig. 7a). The positive correlation between K20/Rb and Ba/Rb and the negative correlation between Rb/Cs and Ba/Rb suggest that,

528

100

c- O ~ " 1 0

O O rr

d

Low-Abundance Lanai Tholeiites Range for Average NMORB

c /

e._ ii

~i!i:::ii:" . . . . " o

�9 Wawaeku (14.8 wt.% MgO) /:~ Manele (7.2 wt.% MgO) �9 Kalamanui (10.9 wt.% MgO) A , ~ i ~ i i i i p i i i

La Ce Pr NdPmSm Eu Gd Tb Dy He Er TmYb Lu

Fig. 4A, B. Rare-earth element abundances in Hawaiian tholeiites normalized to mean Cl-chondrite (Anders and Grevesse 1989). A Representative low-abundance Lanai tholeiites from the Manele, Kalamanui, and Wawaeku stratigraphic sections. The range for averaoe normal mid-ocean ridge basalts (NMORB; Hofmann 1988; Sun and McDonough 1989) is shown by the stippled field. B

100

i m

"O c- O c - 1 0

O

O O

r r

,_~_,_ 'M~sc Hawaiian 'Thoieiite~

�9 Kilauea basalt (7.3 wt.% MgO) [ ] Kilauea picrite (17.5 wt.%MgO) z~ Haleakala pierre (16.6 wt.% MgO) �9 Koolau basalt (11.8 wt.% MgO) x Kahoolawe pierre (16.7 wt.% MgO) B o Mauna Lea picrite (23.7 wt.% MgO)

' ' i . ~ , , i i ~ p i i

La Ce Pr NdPmSrn Eu Gd Tb Dy He Er TmYb Lu

Miscellaneous tholeiitic basalts and picrites from Kilauea (basalt: BHVO-1, Table 1; picrite: BVSP 1981), Mauna Lea (BVSP 1981), Kahoolawe (Leeman et al. in press), Koolau (Roden et al. 1984) and Haleakala (Chcn and Frey 1983). MgO contents listed are based on major element data recalculated to 100% on an anhydrous basis with all iron as F e e

Low-Abundance Lanai Tholeiites

100

c

10 > , m

, m

E

O O re

0.1 O s R b B a T h U N b K L a e e P b P r N d S r S m H f Z r r i E u G d r b D y H o u E r r m Y b L u

Fig. 5. Incompatible trace element abundances in representative low-abundance Lanai tholeiites normalized to primitive mantle. Shown for comparison are a typical Kilauea tholeiite (BHVO-1; data listed in Table 1), the average OIB estimate of Sun and McDonough (1989), and a stippled field representing the range for NMORB (Hofmann 1988; Sun and McDonough 1989). Normali- zing values are from Hofmann (1988). The sequence of elements is based on the order of increasing compatibility in oceanic basalts suggested by Hof~ann (1988)

during surface alteration of Lanai lavas, Rb was more mobile than either K or Cs. Chert and Frey (1985) also concluded that Rb was lost preferentially to K in lavas from Hateakala. In contrast, altered pillow rims of submarine lavas have lower K/Rb, Cs, and K/Cs than do fresh pillow interiors, indicating an apparent increasing order of mobility of K < Rb < Cs (Hart 1969).

Some of the altered Lanai samples have negative Ce anomalies, resulting in relatively high La/Ce ratios compared to those in unaltered Lanai lavas (0,39 0.53 versus 0.34-0.41; Fig. 3@ The negative Ce anomalies in rocks altered by seawater have been attributed to oxidation of Ce to its tetravalent state (e.g., Ludden

and Thompson 1979). Alteration does not, however, appear to have produced systematically lower Ce/Pb ratios in Lanai lavas, as unaltered samples span a larger range in Ce/Pb than do altered samples (12-33 versus 16-26; Fig. 7b).

The scatter of data on plots of U versus other highly incompatible trace elements (e.g., Th, LREE) is probably the result of U loss during subaerial alteration of some of the sampled Lanai lavas. Studies of oceanic basalts have shown that U is mobile during low- temperature alteration (Macdougall et al. 1979; Hart and Staudigel 1982). The U loss in Lanai lavas produced generally higher Nb/U and Th/U ratios in altered relative to unaltered samples (Fig. 7c). Transitional tholeiitie and alkalic lavas from the rainy east flank of Mauna Kea also have high Nb/U and Th/U ratios that are probably the result of surface alteration (Kennedy et al. 1991).

The lack of significant correlation between most incompatible trace elements (e.g., Th, Nb, Zr, Hf, REE, Y, Pb, Ba, Sr) or ratios of

35 OIB + MORB �9

\ \

. Q 13_ 2 5

O

15.

=============================================================================:~:~:~: ,-:.: : , : . :q~, : . : ' . ' . . . . . ' . - : . . . . - : : :+ . . : : . : , : - :+ . , :+ : . : . . - . . . . . . . -: ; ; ;- ; .N,:. : , : . : . ; . :

~ Pr imi t i ve M a n t l e 5 ~ 1 i i i i , i i

0 2 4 6 8 10 12 La

Fig. 6. Ce/Pb versus La for Lanai tholeiites. The range in Ce/Pb for ocean island basalts [OIB] and mid-ocean ridge basalts [MORB] (25 _+ 5; Hofmann et al. 1986; striped field) and for primitive mantle (Hofmann 1988; Sun and McDonough 1989; stippled .field) are shown for comparison

500

400 D ..~_~ 3 0 0

a3200

1 O0

.\

o

A

0.0 1.0 2.0 3.0 K 2 0 / P 2 0 5

500 B

400 �9

~.~.300

m 2 0 0 �9

1 O0 �9

0 . . . . . . . . . . . . .

1 0 1 5 20 25 30 35 Ce/Pb

. . . . . . . . . . . . . . � 9 . . . . , ,

soo C

400 / "

J:3 ~ O ~, 300

Q3200 �9

1 00 �9

25 50 75 1 00 1 25 1 50 Nb/U

Fig. 7A-C. Variations in (A) KzO/P205, (B) Ce/Pb, and (C) Nb/U with Ba/Rb ratios for lavas from Lanai, showing the geochemical effects of subaerial alteration. See text for discussion

529

t-

O

0 (I. o c-

e~

P

[ ]

[]

�9 []

[]

[]

[ ]

[] A

0.5 1.5 2.5 3.5 La/Hf

r

._<2 ( R

O IL r r

Q .

.e-, (D

r

O <..,

O 0. .o_ p .

r

u'J

[ ]

[ ] [ ]

[ ] &

[] B

[ ]

10 15 20 25 30 35 Ce/Pb

[ ]

[ ] �9

[ ]

[ ]

C

0.5 1.0 1.5 2.0 2.5 3.0 3.5 Nb/La

Fig. 8A-C. Incompatible trace element ratios versus relative stratigraphic position for tholeiites from Lanai. A La/Hf, B Ce/Pb, and C Nb/La ratios change progressively upsection within the Manele (triangles; approx. 163 m thick) and Wawaeku (open squares; approx. 43 m thick) sections

those elements (e.g., Nb/Th, Zr/Hf, La/Sm, La/Yb, La/Th, La/Hf, Sr/Nd, Ba/Zr, Ce/Pb) and K20/PzO 5 and Ba/Rb ratios indicates that absolute abundances of these incompatible elements and their ratios were not affected significantly by alteration.

Temporal geochemical variations

Lavas from two of the three sampled stratigraphic sections on Lanai display systematic correlations between relative stratigraphic position and incompatible trace element ratios. At Manele, the thickest sampled section, incompatible trace element ratios change progressively with relative stratigraphic position (Fig. 8). Nb/Th, Nb/U, Nb/La, Sr/Nd, Zr/Sm, and Eu/Eu* ratios systematically decrease upsection and Ce/Pb, La/Hf, La/Th, La/Sm, and La/Yb ratios systematically increase upsection. Lavas from the Wawaeku section, the next thickest sampled section, show similar, although slightly scattered, trends for Nb/La, Sr/Nd, Zr/Sm, Ce/Pb, La/Hf, and La/Th ratios. Lavas from Kalamanui Gulch, the smallest section, show no systematic temporal variations in incompatible trace element ratios. None of the sections show any evidence for systematic temporal variations in major element contents or absolute trace element abundances.

Discussion

Low-abundance Lanai tholeiites

The low-abundance Lanai tholeiites are distinct from other Lanai tholeiites and Hawaiian tholeiites in general. The low-abundance tholeiites contain significantly lower abundances of certain highly incompatible elements (e.g., Th, REE, P, alkalies, Y), are less LREE-enriched (i.e., they have nearly flat REE patterns), and have positive Eu anomalies that are unrelated to their modal plagioclase contents, These geochemical characteristics are a primary feature of the lavas and do not reflect alteration (see the section on alteration). There are several potential mechanisms that could have produced these geochemical features: (1) fractionation of minor accessory minerals, (2) differences in modal mineralogy of the source mantle, (3) interaction with hydrothermal fluids, (4) assimilation of

530

depleted (MORB) wall-rock, (5) accumulation or resorption of phenocrysts or xenocrysts, (6) formation by higher degrees of partial melting than were other Hawaiian tholeiites, assuming the respective sources had similar compositions, or (7) derivation from a less-enriched source containing significantly lower abundances of these elements, assuming the primary magmas of other Hawaiian tholeiites were generated by similar degrees of partial melting. The viability of these mechanisms is discussed next.

Fractionation of accessory phases

Fractionation of mineral phases common to basaltic rocks (i.e., olivine, plagioclase, clinopyroxene) is incapable of depleting these magmas in incompatible elements (e.g., REE, Th, U, Zr, Hf, Y, alkalies, P). However, certain of these elements are compatible in minor accessory phases found in some basalts (e.g., amphibole, apatite, sphene, zircon, phlogopite). Fractionation of such phases could decrease the abundances of some elements, normally con- sidered incompatible, in residual melts.

However, both careful petrographic examination and geochemical evidence have revealed no evidence of any of these accessory phases in Lanai lavas. The strong, positive correlations between highly incompatible trace elements (e.g., Th, LREE) and elements incorporated into potential accessory minerals [P20~-apatite; Zr-zircon, amphibole; Ba-phlogopite; middle rare-earth elements (MREE) sphene, apatite; HREE garnet, spheric, zircon, amphibole] do not support significant fractionation of these phases. Furthermore, the nearly flat to slightly enriched REE patterns of the Lanai tholeiites preclude significant fractional crystallization of either HREE-rich (e.g., garnet, amphibole, zircon) or LREE + MREE-rich (e.g., apatite, sphene, zircon) accessory phases.

Mantle source mineralogy

The retention of minor accessory phases in the mantle during melt extraction can produce magmas containing anomalously low abundances of certain incompatible elements. For example, the high field-strength element (HFSE) (e.g., Zr, Hf, Ti, Nb, Ta, V) depletions in some arc basalts have been attributed to the retention of titanates in the residual source of these lavas following partial melting (e.g., Saunders et al. 1980). However, it is unlikely that the presence of residual accessory phases in the Lanai source produced the low incompatible trace element contents of the low-abundance tholeiites. Such a scenario would require a fortuitous combination of residual accessory phases (e.g., phlogopite, apatite, amphibole, zircon, spheric) such that elements of differing geochemical affin- ity (e.g, Ba, REE, Zr, Y, Th, Nb, Pb) would be comparably low in the resulting magmas.

Could the low-abundance Lanai tholeiites have formed from a source that is identical geochemically to the source of other Hawaiian tholeiites, but which differs in the relative proportion of garnet and clinopyroxene? The melting of 2 sources having identical geochemical compositions but with different clinopyroxene/garnet ratios can produce magmas that contain different absolute REE

abundances and have different REE patterns. For example, at the same percent partial melting, melts formed from a low clinopyroxene/garnet source will have higher HREE concentrations and steeper negative REE patterns than would melts formed from a high clinopyroxene/garnet source. These geochemical differences result because garnet contains higher absolute HREE abundances than does clinopyroxene and enters the melt preferentially over clinopyroxene. However, differences in the proportion of garnet and clinopyroxene in the Lanai source relative to the source of other Hawaiian tholeiites cannot explain the low incompatible element contents of the low-abundance Lanai tholeiites. Concomitant deficiencies both in elements whose concentrations are controlled by residual garnet or clinopyroxene (e.g., HREE, Y) and in elements that are highly incompatible in these phases (e.g., LREE, Th, U) cannot result solely from differences in clinopyroxene/garnet ratios.

Assimilation of MORB wall-rock

The low-abundance Lanai tholeiites have REE, Zr, Hf, and Y abundances lower than average NMORB (Figs. 4-5). Unless the Hawaiian lithosphere is significantly more depleted in these elements than typical NMORB, both direct bulk assimilation or partial assimilation of MORB wall-rock by ascending Lanai magmas are pre- cluded. In addition, isotopic data for Hawaiian shield- buildin,q lavas are inconsistent with a depleted component in their source (West et al. 1987). Isotopic data for Lanai tholeiites indicate these lavas were derived from a source containing a significant "Koolau" component (i.e., radiogenic Sr, unradiogenic Pb) (Fig. 2). This component has been identified as enriched mantle (e.g., Staudigel et al. 1984; Zindler and Hart 1986; Sun and McDonough 1989) and primitive (e.g., White 1985; West and Leeman 1987) mantle, but it is inconsistent with a depleted MORB source or depleted MORB lithosphere.

Hydrothermal alteration

Although submarine hydrothermal waters are character- ized by pronounced positive Eu anomalies (e.g., Michard et al. 1983; Campbell et al. 1988), interaction between such fluids and stored or ascending magmas could not, re- alistically, have produced the low incompatible element abundances and positive Eu anomalies in the low-abundance Lanai tholeiites. First, REE concentrations in hydrothermal fluids are too low (10 - 9 g_g-1 to 10 -12 g_g-1 range, typically) to have affected the geochemistry of these lavas (ppm range) significantly, except at unrealistically high water/rock ratios (e.g., W/R > 105; Michard and Albarede 1986). Second, hydrothermal waters have steep LREE-enriched patterns, which are uncharacteristic of the low-abundance Lanai tholeiites. Third, the Lanai samples from this study reveal no petrographic signs of the extensive hydrothermal alteration that such high water/rock ratios imply, although, admittedly, the affects of very minor hydrothermal alteration may not be obvious.

Phenocryst accumulation and assimilation

Phenocryst accumulation can produce low abundances of incompatible elements in magmas via a dilution effect. However, the principal phenocryst phase in Lanai tholeiites is olivine whose accumulation cannot account for the substantial differences in La/Sm and La/Yb ratios and in absolute REE abundances observed for lavas with equivalent MgO contents (Fig. 3b).

The low-abundance Lanai tholeiites have positive Eu anomalies whose magnitudes generally increase with decreasing incompatible element abundance. The most common interpretation of positive Eu anomalies in basaltic rocks is plagioclase accumulation or assimilation. How- ever, there are several lines of evidence which taken in toto negate these processes as the principal cause of the Eu anomalies found in Lanai tholeiites. 1. The Eu/Eu* ratios of Lanai lavas are not correlated with modal plagioclase; therefore, it is unlikely that the positive Eu anomalies result from plagioclase phenocryst accumulation. Most Lanai lavas lack relict grains and embayed phenocrysts of plagioclase. Thus, if plagioclase assimilation was involved in the genesis of these lavas, it would have to have been exceedingly efficient (i.e., plagioclase would have to have been completely resorbed), as has been suggested by Elthon (1984) for aphyric oceanic basalts. 2. The major and trace element compositions of Lanai tholeiites are inconsistent with these lavas having under. gone significant plagioclase fractionation or assimilation. The major element compositions of Hawaiian tholeiites with MgO > 6.8 wt.% are controlled primarily by olivine subtraction and addition (e.g., Wright 1971). Like tholeiites from other Hawaiian volcanoes, lavas from Lanai have relatively constant AlzO3/CaO with decreasing MgO and increasing FeO*/MgO, up to FeO*/MgO = 1.8, where, presumably, pyroxcne and plagioclase join olivine as fractionating phases (Fig. 9). This trend is inconsistent with

1.8 Mauna Loa ~ Koolau

(~1.4

1.0 ~ ' ' ' 0.5 ~.0 ~.5 2.0 2.5

FeO*/MgO

Fig. 9. AI203/CaO versus FeO*/MgO for tholeiites from Lanai (triangles), where FeO* is total iron recalculated as FeO. Fields for Kilauea, Mauna Loa, and Koolau tholeiites are shown for comparison. See text for discussion. Data sources: Garcia (unpub.), Macdon- ald and Katsura (1961, 1964). Macdonald (1968), Jackson and Wright (1970), Wright (1971), Wright and Fiske (1971), Wright et al. (1975), Moore et al. (1980), BVSP (1981), Roden et al. (1984), Thompson et al. (1984), Lipman et al. (1985), Newsom et al. (1986), Lockwood et al. (1987), Ho and Garcia (1988), Wilkinson and Hensel (1988)

531

1.7

1.6

1.5

5t.4

LU 1.2

1.1

1.0

0.9 5.5

"-I

LU

A

�9 �9 & �9

k �9 �9

1.7

6.0 6.5 7.0 7.5 8.0

A I 2 O a / T i O 2

1.6

1.5

1.4

1.3

1.2

1.1

1.0

0.9

B

&�9 �9

AA AA A �9

0.5 1.0 1.5 2.0 2.5 3.0

Nb/ka

8.5

3.5

60

50

40 Z

30

20

10 0.5

C ,,

i A

A,O.

1.0 1.5 2.0 2.5 3.0 3.5

N b / L a

Fig. 10A C. Correlations between major and trace element ratios sensitive and insensitive to plagioclase addition or removal for Lanai lavas. A Eu/Eu* is not correlated with A1203/TiO 2 ratios, indicating that plagioclase addition or removal was not significant in these lavas. B and C Highly incompatible trace element ratios insensitive to plagioclase control (Nb/La) are positively correlated with ratios indicative of plagioclase fractionation or accumulation (Eu/Eu* and Sr/Nd)

either plagioclase fractionation or assimilation in these lavas.

Plagioclase assimilation should produce readily observable effects on the abundances of certain elements and on certain elemental ratios sensitive to plagioclase (e.g., A1203, A1203/CaO , AIzO3/TiO2, Sr). However, Eu/Eu* ratios in Lanai lavas do not correlate with these geochemical parameters, but do correlate strongly with trace element ratios insensitive to plagioclase addition or removal (e.g., Nb/La, Nb/Th, La/Th) (Fig. 10). Therefore, we do not interpret correlations of ratios such as Sr/Nd and Eu/Eu* with Nb/La, La/Th, Nb/Th, and Ce/Pb as having been produced by plagioclase assimilation. 3. We quantitatively tested the possibility that Lanai magmas containing higher absolute REE abundances and no Eu anomalies (potential parent magma) could have

532

assimilated plagioclase to produce magmas containing lower REE abundances and positive Eu anomalies (potential daughter magma). First, pairs of potential parent and daughter magmas were selected from each of the sampled stratigraphic sections (potential parental magmas: LMan-5, LKal-5, LWaw-5; potential daughter magmas: LMan-2, LKal-4, LWaw-1). Then, we calculated the REE composition of plagioclase required to produce the observed daughter magmas through assimilation of plagioclase by the potential parent magmas. The results for one example (the LMan-5/LMan-2 parent/daughter pair) are shown in Fig. 11. For this example, no solutions are obtained for < 70% plagioclase assimilation. Other potential parent-daughter magmas produce similar results: the LKal-5/LKal-4 and the LWaw-5/LWaw-1 pairs require > 80% and > 60% plagioclase assimilation, respectively. Calculated plagioclase REE compositions range from nearly flat to LREE-depleted, i.e., (La/Sm)cN _< 1. In contrast, REE patterns of plagioclase megacrysts and phenocrysts reported in the literat- ure typically are significantly more LREE-enriched and have more pronounced positive Eu anomalies (Fig. l l insert).

The results of REE modeling calculations show that implausible amounts of plagioclase assimilation ( > 60%) are required to produce the positive Eu anomalies of the

i i

L M a n - 5

1 00

0

r-- L M a n - 2 0 a \ aJa~_a~ ( ' - 1 o O 90o - . . . . .

0 7 r r

Calculated Plagioclase i i i i i i ~ i I i i

La Ce Pr Nd PmSm Eu Gd Tb Dy He Er TmYb Lu

~) 100

i-- " 0 C 0 r-" 10

o

0 1 0

rr

0,1

La Ca Pr NdPmSm Eu Gd Tb Dy He Er Tm Yb Lu

Fig. 11. Calculated rare-earth element compositions of plagioclases required to produce low-abundance Lanai tholeiite LMan-2 through assimilation of plagioclase by a potential parent magma having the composition of sample LMan-5. Calculated plagioclase compositions for 70-90% assimilation are shown as dashed lines. No solutions are obtained for < 70% plagioclase assimilation. Insert shows REE compositions of representative plagioclase phenocrysts and megacrysts (Higuchi and Nagasawa 1969; Schnetzler and Phil- potts 1970; Paster et al. 1974; Fujimaki et al. 1984; Morse and Nolan 1985; Phinney and Morrison 1990)

low-abundance Lanai tholeiites (Fig. 11). Such large amounts of plagioclase assimilation would drastically increase the AI20 3 and Sr contents of the resulting magmas. However, the low-abundance Lanai tholeiites have lower or just marginally higher A1/O 3 and Sr contents than do associated lavas containing higher incompatible element abundances and no Eu anomalies. Using the calculated equilibrium plagioclase composition of LMan-5 (Anvv), 60% plagioclase assimilation would produce a magma containing substantially higher A120 3 (21.8 versus 13.9 wt.% AlzO3) and Sr (434-556 versus 330 ppm Sr; based on Sr abundances in plagioclase separates from basalts; Philpotts and Schnetzler 1970; Fujimaki et al. 1984) contents than those observed for LMan-2. Even for plagioclase assimilation as low as 10%, the resulting magma would have higher A120 3 and Sr contents (14.7 wt.% A1203, 348-369 ppm Sr) than LMan-2. The other models tested produced similar results.

We conclude that the low-abundance tholeiites cannot be related to other Lanai lavas by plagioclase assimilation or accumulation.

Source geochemical heterogeneity and partial melting variations.

On the basis of the arguments discussed, we reject fractionation of minor accessory phases, differences in mantle source mineralogy, hydrothermal alteration, assimilation of depleted (MORB) wall-rock, and phenocryst accumulation or assimilation as mechanisms capable of producing the geochemical features of the low-abundance Lanai tholeiites. In the following sections, we examine whether the distinctive characteristics of these lavas resulted from variable degrees of partial melting, geochemical heterogeneity in the Lanai source, or a combination of these processes. 1. Lanai primary magmas. To evaluate the roles of partial melting and source heterogeneity in the genesis of Lanai lavas, primary magma compositions must be inferred. Estimates of Hawaiian primary magma compositions range from 12~5 wt.% MgO (e.g., Wright 1971, 1984; Irvine 1977; Maaloe 1979; Leeman et al. 1980; BVSP 1981; Chen and Frey 1983; Wilkinson 1985; Nicholls and Stout 1988; Wilkinson and Hensel 1988; Garcia and Hulsebosch 1990; Clague et al. 1991; see Table 4). The most mafic Lanai lava in our sample suite is picritic (sample LWaw- 4b: 16 vol.% olivine phenocrysts, Mg4~ ~ 71), but it is unlikely that it represents a primary magma. Petro- graphic examination reveals that many of the olivines in LWaw-4b are xenocrysts, which suggests that it is accu- mulative. The high Ni/MgO ratio of this sample (Ni/MgO = 48.4) is also indicative of olivine accumulation (BVSP 1981). Furthermore, LWaw-4b contains plagioclase and clinopyroxene phenocrysts, which in Hawaiian tholeiites .usually form only in evolved liquids ( < 7 wt.% MgO; Wright 1971).

Our best sample for inferring the composition of Lanai primary magmas is LKal-4. It has nearly 11 wt.% MgO, Ni/MgO = 26.3, approx. 5 vol.% olivine phenocrysts, no plagioclase or elinopyroxene phenoerysts, and the lowest abundances of several incompatible elements (e.g., REE, P, Hf, Y) among samples from this study. We infer that LKal-4 probably evolved from primary magma by olivine

Table 4. Major and trace element data for estimated Hawaiian primary element data are given in ppm

533

magmas. Major element data are given in weight percent. Trace

1 2 3 4 5 6 7 8 9

Si02 46.75 47.95 48.05 48.37 46.95 46.17 48.00 48.74 49.57 A1203 9.01 9.29 9.89 11.36 13.10 7.83 11.00 11.10 10.51 F%O3 - 13.44 1.14 - - 2.27 2.06 FeO 11.13 10.80 10.25 11.00 11.36 11.00 9.23 9.41 MgO 21.81 21.19 16.31 14.07 14.55 25.00 14.50 15.51 17.03 CaO 7.30 7.03 7.77 9.58 10.16 6.29 9.00 8.37 7.66 Na/O 1.47 1.51 1.67 1.90 1.73 1.26 2.00 1.90 1.99 K20 0.31 0.23 0.26 0.49 0.08 0.23 0.40 0.37 0.15 PzOs 0.16 0.14 0.16 0.23 - 0.12 - 0.20 0.08 TiO2 1.58 1.37 1.83 2.28 2.02 1.35 1.35 1.71 1.42 MnO 0.16 0.19 0.17 0.18 0.15 0.22 - 0.16 0.12 CrzO 3 0.21 0.20 0.19 - 0.12 NiO - - 0 . 0 9 . . . . 0.09 -

La 8.2 5.5 7.1 . . . . . 2.20 Ce 22.0 14.7 18.8 - - 5.89 Nd - 12.3 - - - 4.2 Sm 3.7 3.2 3 . 7 3 . . . . 1.34 Eu 1.15 1.06 1 . 3 9 . . . . 0.63 Tb 0.55 0.55 0.7i . . . . 0.248 Ho - - 0.8 0.277 Yb 1.22 1.3 1 . 7 3 . . . . . 0.70 Lu 0.24 - - - 0.098

1, Calculated Kilauea primary magma (Wright 1984) 2, Calculated Mauna Loa primary magma (Wright 1984) 3, Postulated primary magma, Haleakala tholeiite (Sample C-122; Chen and Frey 1983) 4, Kilauea Iki primitive liquid (Irvine 1977) 5, Kilauea Iki parent magma (Green and Ringwood 1967) 6, Prehistoric Kilauea caldera primary magma assuming 25 wt.% MgO (Wright 1971) 7, Hawaiian parent magma (BVSP 1981) 8, Mauna Loa picrite primary magma (Wilkinson 1985) 9, Calculated Lanai primary magma for sample LKal-4. See text for method of calculation. Major element data normalized to 100% after

setting FeZ+/(Fe 2+ + Fe 3+) to 0.90

fraetionation alone. Maaloe and Hansen (1982), Garcia and Hulsebosch (1990), and Clague et al. (1991) proposed that some Hawaiian pr imary tholeiites have M g O contents of approx. 17 wt.%. A mixture of 85% LKal-4 and 15% Fo90 olivine would yield a m a g m a with 17 wt.% MgO. The choice of olivine composi t ion has little affect on the calculated pr imary m a g m a composit ion. For example, a less-forsteritic olivine (e.g., Fo86.8, the calculated equilibrium olivine composi t ion for LKal-4) changes the estimated mixing propor t ions by just 1%. To calculate the REE composi t ion of this Lanai pr imary magma, the REE contents of LKal-4 were diluted through the addit ion of 15% olivine phenocrysts (Table 4, column 9). The calculated REE abundances for this Lanai pr imary magma, as well as abundances in sample LKal-4 itself, are lower than those calculated by Wright 0984) or Chen and Frey (1983) for Kilauea, M a u n a Loa, and Haleakala pr imary magmas (Table 4).

The calculated Lanai pr imary m a g m a and other calculated modera te -Mg Hawaiian pr imary magma compositions (14 17 wt.% MgO) fall along the 2 0 k b cotectic defined by glasses f rom the reversal experiments of Fa l loon et al. (1988) on calculated equilibrium partial melts of Hawai ian pyrolite (Fig. 12). This pressure is within the stability field of garnet, depending on temperature and volatile content (e.g., O ' H a r a et al. 1971; Green 1973; Maaloe and Jakobsson 1980; Taylor and Green

1988) and corresponds to approx. 65 km depth. Earth- quakes occur under the active Hawai ian volcanoes to depths of approx. 60 km (Eaton and Mura ta 1960; Eaton t962; Ryan et al. 1981; Dzurisin ct al. 1984; Klein et al. 1987). Sixty to sixty-five kilometers is a min imum depth for segregation of Hawai ian tholeiitic magmas from their source. The source of these magmas must lie below the lithosphere because thermodynamic constraints (Morse 1991), volume considerations (Frey and Roden 1987), and isotopic composi t ions (West et al. 1987) preclude Ha- waiian tholeiites from being derived th rough melting of oceanic lithosphere. Estimates of the depth to the base of the lithosphere beneath Hawaii range from about 70 km to > 100 km (e.g., Leeds 1975; Wat ts 1978; Liu and Chase 1989). 2. Partial melting models. Experimental, major element, and REE data yield conflicting interpretations concerning the amount of melting required to generate Hawai ian pr imary tholeiites (see Feigenson 1986; Frey and Roden 1987; Frey et al. 1990b). Partial melting estimates based on geochemical models range from 2%, leaving a garnet lherzolite residue (Budahn and Schmitt 1985), to 40%, leaving a harzburgite residue (Wright 1984). The low degrees of melting suggested by Budahn and Schmitt (1985) require unrealistically large volumes of sub- lithospheric mantle to be consumed. For a shield the volume of Lanai and a postulated plume diameter of

534

10 kb<

20kb 1 5 i / / 30 kb

0] Hy

�9 Lana i Tholeiites * Lanai P r i m a r y Magma

Hawai ian P y r o l i t e �9 Hawai ian P r i m a r y Magmas

q

Fig. 12. Olivine (Ol) Quartz (Q) Plagioclase (Pl) molecular nor- mative ternary comparing Hawaiian primary magma estimates, including Lanai, with olivine + orthopyroxene + clinopyroxene _ spinel + liquid cotectics of Hawaiian pyrolite at 5, 10, 15, 20, and 30 kb (cotectics from Falloon et al. 1988; Jaques and Green 1980). The locus of liquids in equilibrium with a dunitic residue is defined by the tie-line intersecting the olivine apex and Hawaiian pyrolite. Moderate-MgO (14 17wt.% MgO) Hawaiian primary magmas (solid dots)are given in Table 4. All data were normalized to 100% on an anhydrous basis after setting Fe2+/(Fe 2+ + Fe 3+) to 0.90

25 km (approx. the areal diameter of Lanai), a cylindrical volume of sub-lithospheric mantle more than 1200 km deep would have to be melted. This is a minimum estimate, because the volume of the Lanai shield calculated by Bargar and Jackson (1974) may be up to 50% too low because it ignores subsidence (Moore 1987). Wright's (1984) model also seems to require an excessive volume of sub-lithospheric mantle to be melted (2.5 x 106 km 3) to account for the postulated volume of enriched material (25 000 km 3) (see Frey and Roden 1987). This is equivalent to a cylindrical volume of sub-lithospheric mantle approx. 5100 km deep.

To constrain the extent of partial melting in the Lanai source, we evaluated several melting models using our calculated primary magma composition. A primitive source (i.e., flat REE pattern, REE abundances approx. 2.6x chondrite; Hofmann 1988) was used, because Lanai lavas have Sr, Pb, and Nd isotopic compositions possibly consistent with such a component in their source. The maximum amount of partial melting required to generate the Lanai primary magma from a primitive source can be calculated from the equilibrium non-modal melting equa- tion (Shaw 1970) assuming bulk D = 0 for highly incompatible trace elements (e.g., LREE). This approach yields values in the range 25-29% partial melting for LREE to MREE (La ~ Sm). Using bulk D values calculated from relevant mineral/melt D LREE values (Prinzhofer and Alle- gre 1985) for a source containing 5 14% garnet, the

maximum amount of partial melting is constrained to 26-28% partial melting.

We calculated melt compositions for up to 30% modal and non-modal equilibrium batch melting for a range of garnet peridotite source modal mineralogies; a representative result is shown in Fig. 13a, b. Calculated melting curves do not match REE patterns of the Lanai primary magma or the low-abundance tholeiites. Although LREE abundances in the Lanai primary magma are consistent with approx. 25% partial melting, calculated HREE abundances do not match those observed (Fig. 13a). The REE patterns of Kahoolawe and Mauna Loa picritic tholeiites fall roughly along the 10% melting curve, although HREE abundances do not match closely (Fig. 13b). If these Hawaiian picrites are not primary magmas, then the degree of partial melting must be > 10%, pro- vided these lavas are not cumulates. Similar melting calculations were performed for a spinel lherzolite source, but neither modal nor non-modal melting curves could repro- duce the observed enrichments in LREE over HREE for either the Lanai primary magma or the Hawaiian picritic tholeiites. To yield the observed REE compositions of these potential primary magmas, a spinel lherzolite source must be LREE-enriched.

Several incremental melting models were tested; results for the most successful one are shown in Fig. 13c. The LREE (La --* Sm) and HREE (Tm ~ Lu) contents of the Lanai primary magma fall approximately along the calculated melting curve for 30% non-modal melting of a pyrolite source. Calculated and observed MREE (Gd ~ Er) contents differ significantly, which could be an artifact of the non-linear increase in published MREE HREE distribution coefficients for garnet. Rare-earth element distribution coefficients for clinopyroxene and garnet in mantle peridotite under high pressure and temperature are not well-determined. Thus, incremental melting potentially can account for the relative LREE-enrichment of the Lanai primary magma without having to invoke a LREE-enriched source. However, this mismatch also could result if the Lanai source is more LREE-enriched (i.e., higher LREE/HREE) than the source composition used in the modeling.

We calculated potential source REE compositions for various degrees of partial melting of a garnet peridotite source containing 5-14% garnet, leaving garnet as a residual phase. If the Lanai primary magma formed by _> 20% partial melting of such a source, then the Lanai

source was LREE-enriched. For 25% partial melting, roughly consistent with maximum partial melting estimates, the Lanai source would have contained near- primitive absolute LREE abundances (2.5x chondrite) and 1.1x to 1.8x chondrite HREE abundances (Fig. 13d).

Isotope systematics show that the source of Hawaiian shield-building lavas is heterogeneous (e.g., Staudigel et al. 1984; Tatsumoto et al. 1987; West et al. 1987). Therefore, the calculated REE composition of the Lanai source must represent a mixture. If the Hawaiian source is a mixture of primitive and enriched components, we can constrain their proportions in Lanai magmas using calculated REE compositions of the bulk Lanai source (Fig. 14). For a mixed source consisting of > 40% primitive component, no solutions are obtained (i.e., negative REE abundances

535

100:

10 �9

�9 1 x= ~DIO0

r O

10:

. 01

. 5 0 - - -

a Manele (7.3 wt.~ MgO) [ ]Kalamanui (11.0 wt.~ MgO) ~r Kalgmanui Primqry Magma

(17 wt.~ MgO) A

.0~_ B a t c h M e l t i n g " ' - - - Garnet LheTzoli~e

" " - . Model o , . ~ 0:10:10

a Kilauea picrite (19.! wt.~ MgO) . A Mauna Loa pierite (21.2 wt.~, MgO,) ~-Kahoolawe picrite ,(16.7 wt.~ MgO) B O Haleakala picrite (16.6 wt.~ MgO

La Ce Pr Nd PmSmEu Gd Tb Dy Ha Er TmYb Lu

/

B a t c h M e l t i n g | . o , . . . I n c r e m e n t a l M e l t i n g Garnet Lherzoli~e " ' , , . Garnet Lherzolite

Model " ' - . . Model 60:20:10:10 " '- 61:15:10:14

. 10 . . . . . . ~ ~ - -~_ "~

oWawaeku (15.0 wt.~ M gO) zx Uanele (7.3 wt.~ MgO) u Kalamanui (11.0 wt.~ MgO) #Kalgmanui Primary Magma C

(17 wt.~ MgO)

S o u r c e C o m p o s i t i o n s P r i m a r y Magma F = .25

Primitive Mantle . ~ - a ~ . 4 . . . . ~ ;~ , , o ~ o - o - o

zx Model 60:20:10:10 - ~ . . a . , ~ _ o 4 .

�9 Model 61:15:10:14 " ~ - - a ' - a - - ~ . . ~

Lo Ce Pr Nd PmSmEu Gd Tb Dy Ha Er TmYb Lu

10

Fig. 13A D. C1 chondrite-normalized rare-earth element plots for selected partial melting models. A Calculated non-modal equilibrium batch melting curves (dashed lines) for 1%, 10%, 20%, and 30% non-modal partial melting of a primitive mantle source composition (data from Hofmann 1988). Assumed source modal mineralogy: olivine 60%, orthopyroxene 20%, clinopyroxene 10%, and garnet 10%o (based on modal abundances in the undepleted, sheared garnet lherzolite PHN-1611; Boyd and Nixon 1972; Shimizu 1975). Modes of minerals entering the melt: olivine 10%o, orthopyroxene 10%, clinopyroxene 40%, garnet 40% (Leeman et al. 1980; Frey 1984). Partition coefficients are from Prinzhofer and Allegre (1985). Shown are representative low-abundance tholeiites from Lanai and

the calculated Lanai primary magma. B Representative tholeiitic picrites from Kilauea, Mauna Loa, Kahoolawe and Haleakala. C Calculated non-modal incremental melting curves (dashed lines) for 1%, 10%, 20%, and 30% partial melting of a pyrolite source (61%o olivine, 15% orthopyroxene, 10% clinopyroxene, 14% garnet; Duffy and Anderson 1989). The melting increment used is 0.1%; curves represent total aggregate melt compositions. D calculated Lanai source compositions (dashed lines) for 25% non-modal batch melting of two source modal mineralogies assuming a primitive mantle source composition. Garnet was not completely consumed during melting

E n r i c h e d Source C o m p o n e n t @

" ~ 10

�9 xS 0 1

o �9

~:~ 0.1

Primary Magma

Primitive Mantle * ~'~'-~'r s 3 o ~~~-~'~'~~ Lanai Source

" , ' , . . . . - - L L - - - .10/.9o

" .~o/.7o

~x

M o d e l 6 0 : 2 0 1 0 : 1 0 ' , \

F .... .2 5 ' .40/.60 . . . . . . . . . , , , , , 7 - - - ,

Lc Ce Pr Nd PmSmEu Od Tb Dy Ha Er TmYb Lu

Fig. 14. C1 chondrite-normalized rare-earth element plot of calculated compositions for the enriched mantle component in the Lanai source for 25% partial melting. The assumed Lanai source is a mixture of primitive and enriched components (see text for discussion of assumptions). Primitive mantle abundances are from Hofmann (1988). Source mode is the same as in Fig. 13a. Enriched component compositions are depicted by dashed lines, with proportions of primitive (PM) and enriched (EM) mantle components in the mixed source indicated next to each line

result). A 40% primitive 60% enriched source would have (La/Yb)r ~ 11.6 for this model. If the primitive component constitutes more than about 5% of the mixed source, then the enriched component has absolute H R E E abundances less than chondritic and contains lower absolute REE abundances than the primitive component.

Variations in partial melting versus source differences

Variations in incompatible trace element ratios in erupted lavas can be produced by extensive crystal fractionation of phases in which these elements have different distribution coefficients (e.g., clinopyroxene), melting a homogeneous source to varying extents, or melting a heterogeneous source. We rule out crystal fractionation as having played a significant role in producing such variations in Lanai tholeiites because olivine was the dominant fractionating phase. Absolute abundances of incompatible trace elements (e.g., REE, Th, U, Pb, Nb, Zr, Hf, St, Ba) in olivine are too low to produce substantial variations in ratios of these elements by fractional crystallization alone. Also, Lanai tholeiites of equivalent MgO content display substantial differences in absolute REE abundances and the

536

degree of LREE/HREE enrichment. For example, three Manele lavas with equivalent MgO contents (7.2-7.3 wt.%) have La/Yb ratios ranging from 2.80 to 5.24 and La and Yb abundances ranging from 2.6 to 9.1 ppm and 0.94 to 1.74 ppm, respectively (Fig. 3b).

Clinopyroxene is the only phase in these rocks capable of modifying La/Yb ratios to such an extent by crystal fractionation. However, geochemical evidence indicates that clinopyroxene fractionation in these lavas was insig- nificant: (1) the positive correlation between Sc and incompatible element abundances (e.g., Hf, Zr, Pb), (2) the somewhat scattered but negative correlation between MgO contents and Sc abundances, and (3) the lack of correlation between MgO contents and AI203/CaO ratios.

If the Lanai source was homogeneous geochemically, then variations in LREE abundances and LREE/HREE ratios for spatially associated lavas with the same MgO content suggest they are related by variable degrees of partial melting. We tested quantitatively whether or not the range in highly incompatible trace element ratios found in Lanai tholeiites is too large to result solely from partial melting variations of a geochemically homogeneous source. Thorimn, niobium, and lanthanum are similar geochemically in that all are highly incompatible during partial melting of typical mantle peridotite. If we assume that the range in Nb/Th for Manele lavas (Nb/ThLM,n_ 6 = 16.0, Nb/ThL~a,,.2 = 33.9) results solely from variations in the degree of partial melting of a homogeneous source and that the Nb/Th ratio of the Lanai source can be estimated, then the percent partial melting required to generate this range can be calculated. These two samples are ideal for testing this hypothesis, because they have equivalent MgO contents (MgOLM,n. 2 = 7.17 wt.%, MgOLM,,-6 = 7.22 wt.%) and therefore appear to represent the same level of fractionation.

First, we assumed that bulk D rh < bulk D Nb, because (1) on a plot of Nb versus Th, Lanai data intersect the Nb axis, (2) Nb/Th is negatively correlated with incompatible element abundances, and (3) this relative order of elemental compatibility in potential mantle phases (olivine, orthopyroxene, clinopyroxene, garnet, spinel) is favored overwhelmingly in published geochemical studies (e.g., Clague and Frey 1982; LeRoex and Erlank 1982; Baxter et al. 1985; Chen and Frey 1985). Second, we assumed that LMan-2 represents a larger degree of melting than does LMan-6, because it contains lower incompatible element abundances and has lower La/Sm and La/Yb ratios. Thus, Nb/ThLian. 2 must more closely approximate the Nb/Th ratio of the source than does Nb/ThL~,._ 6 because, with decreasing amounts of partial melting, trace elements will increasingly become fractionated from one another as a function of the difference in partition coefficients. Using Nb/ThLu,._ 2 for the postulated homogeneous Lanai source, we calculated the percent partial melting required to generate Nb/ThLMa,_ 6. Bulk D Nb and bulk D Th values were calculated over a range of mineral-liquid partition coefficients for both garnet and spinel peridotite assem- blages.

Results of our modeling indicate that 4.8-7.6% partial melting is required to produce a magma with Nb/ThLuan_ 6

from a source with Nb/ThLMan.2. Similar calculations modeling the range in La/Th ratios for Manele lavas (La/ThLMa,_ 2 = 10.9, La/ThLMa,_ 6 = 17.3) require even smaller degrees of partial melting (0.4-1.9%). Note that these partial estimates represent maximum values, because we assume that LMan-2 has Nb/Th and La/Th ratios equal to the Lanai source. If the Lanai source has more extreme Nb/Th or La/Th ratios than does LMan-2, which is likely because this sample must itself be derived from a partial melt, then the percent partial melting necessary to fractionate these ratios to the extent observed will be even lower than our calculated values. Such small degrees of partial melting are probably unrealistic for Hawaiian tholeiites (e.g., Frey and Roden 1987).

Two other observations argue against Lanai tholeiites having been derived solely by variable degrees of partial melting of a geochemically homogeneous source. First, the increase in La/Th ratios with increasing abundances of these elements would require that DLa< D vh, which is unlikely (see Hofmann 1988; Sun and McDonough 1989). Second, variable degrees of partial melting of a homogeneous source, assuming clinopyroxene and/or garnet re- main as residual phases, should produce crossing REE patterns, which are not observed in these lavas.

We conclude that the range in highly incompatible trace element ratios found in Lanai tholeiites is too large to result solely from partial melting variations but must reflect geochemical heterogeneities in the Lanai source. The relative proportions of source components incorporated into primary melts are likely a function of the extent of fusion, assuming these components have different solidi. Geochemical differences in lava compositions between Hawaiian shields could be controlled largely by differences in the proportions of these components in the plume material supplying each volcano and in the extent of partial melting. Lavas from isotopically similar volcanoes may, thus, contain similar proportions of source components, assuming these proportions are controlled by partial melting. If so, then isotopically similar tholeiites from Kahoolawe, Koolau, and Lanai, may have formed by similar amounts of partial melting, in which case the Lanai source contains significantly lower abundances of incompatible elements than the Koolau or Kahoolawe sources. A caveat to such an interpretation should be noted: the large range in Sr-Pb-Nd isotope data for Kahoolawe (West et al. 1987) could reflect significant variations in the extent of partial melting in the source of those lavas.

Considering these complexities, differences in both geochemical composition and extent of partial melting may characterize the respective sources of even isotopically similar tholeiites from different volcanoes. There is no reason to presuppose that the extent of partial melting was invariant over the duration of shield-building activity for any Hawaiian volcano. If each volcano is fed by a discrete blob of plume material derived from the Hawaiian hotspot source and if this source consists of dispersed heterogeneities, then each blob may contain different proportions of such materials. Alternatively, all volcanoes may tap a continuous flux of plume material in which the proportions of isotopically distinct components vary with time.

537

Nature of geochemical heterogeneity in source of Hawaiian tholeiites

Results from the previous section show that the source of Lanai tholeiites is heterogeneous geochemically. Because incompatible trace element ratios in these lavas span relatively large ranges, we infer that the extent of partial melting in their source must have been insufficient to produce geochemically homogenized melts. Plots of incompatible trace elements ratios (e.g., La/Th-Ce/Pb, Nb/Th-La/Hf, Ce/Pb-Nb/Th, Ce/Pb-Nb/La, Ba/La- Zr/Th, Nb/La-Zr/Th) form coherent trends, consistent with mixing between geochemically distinct source components. The low-abundance tholeiites fall at one extreme on these trends and must therefore approximate one of the Lanai source components.

Correlations between highly incompatible trace element ratios and absolute abundances of these elements in Lanai tholeiites (e.g., Figs. 6, 10c) could result from variations in the degree of partial melting of a geochemically heterogeneous source. Because the low-abundance tholeiites fall at one extreme on these trends, they could represent one extreme in the extent of partial melting in the Lanai source. Lower absolute incompatible element abundances and lower La/Sm and La/Yb ratios in the low-abundance tholeiites suggest they represent higher degrees of partial melting than do other Lanai lavas. Thus, this component appears to dominate melt compositions at higher degrees of partial melting, whereas the component represented by Lanai lavas containing higher incompatible element abundances dominates at lower degrees of partial melting. The low-abundance tholeiites could therefore represent a more refractory component (i.e., a less- fusible, higher solidus component) than the other principal component in the Lanai source.

Trace element patterns of the low-abundance tholeiites differ significantly from Lanai lavas containing higher abundances of incompatible elements (Fig. 15). Compared to other Lanai lavas, they are enriched in HFSE (Zr, Hf, Ti, and, to lesser extent, Nb), Sr, and Eu, relative to other incompatible trace elements (e.g., Th, U, REE, alkalies). The low-abundance tholeiites have the lowest Ce/Pb, La/Zr, and La/Hf ratios and the highest Nb/La and Nb/Th ratios among Lanai lavas. Although Ce/Pb ratios of these lavas approach primitive values (Fig. 6), the low- abundance tholeiites have non-chondritic Nb/La and Nb/Th ratios. Therefore, it is unlikely the low-Ce/Pb component represents a geochemically unmodified primitive mantle component.

The Lanai low-Ce/Pb source component is character- ized by low REE/HFSE, Th/HFSE, and alkali/HFSE ratios, indicating it is relatively enriched in HFSE. Con- versely, many arc lavas are enriched in LREE, alkali metals, and alkaline earths, relative to HFSE (e.g., Gill 1981; Dupuy et al. 1982). This so-called depletion in HFSE for arc lavas could result from the retention of residual titanate phases (e.g., rutile, sphene) in their source following dehydration or melting of subducting slab materials (e.g., Saunders et al. 1980). If this residual slab material becomes entrained within the convective regime of the mantle, then it could contribute significantly to the source of oceanic island basalts (e.g., Saunders et al. 1991; Weaver

13l r

. n

E "r-"

13.

O O

r r

10

i ~ i E i i i i i k i i i i i L i i i i i i i i i

Manele BHV 04 / ~ - Low-abundance

/e e.l~i - ~ e ~ tholeiitos

/ ~ / * LMan-2 - - = ~ /l o LMan-3

. . . . . . . . . . . . . . . . . . .

e s T h P b K R b U S r B a N b L a O e P r N d Z r P H f S r n T i E u G d T b D y Y H o E r T m Y b L u

r-.- c~

E EL

O O

s

10

i i i i i i i i ~ i i J i ~ i i i i i i i i i i

Manele BH~4 e/P~-. ~ High-abundance

/ o ~ " ~ tho/eiites

�9 BHVO - ~ �9 LMan-5 0 LMan-6

1 ~ i i i i i i i i i i i i I i i i i i i i i r

C s T h P b K R b U S r B a N b L a e e P r N d Z r P H f S m T i E u G d T b D y Y H o E r T m Y b L u

Fig. 15A, B. Trace element abundances of low-abundance and high- abundance Lanai tholeiites from the Manete section, normalized to primitive mantle (Hofmann 1988). The sequence of elements is based on the order of normalized elemental abundances observed for a typical Kilauea tholeiite: U.S.G.S. standard rock, BHVO-1 (Govindaraju 1989). Trace element patterns for the high-abundance Lanai tholeiites resemble that of BHVO-I, with a few exceptions (notably, K and Sr). In contrast, the low-abundance Lanai tholeiites have elevated HFSE contents relative to other incompatible elements (e.g, REE, Th, U, alkalies)

1991). Because subducted residual slab material has undergone previous melting and/or dehydration, it is likely to be refractory (McDonough 1991), possibly more so than the surrounding mantle material in which it becomes entrained. Thus, the geochemical signature im- parted to generated melts by such a component residing in a mixed source may be evident in the chemistry of erupted lavas only at relatively large degrees of partial melting. The apparent relative enrichment in HFSE abundances in the low-abundance Lanai tholeiites may reflect the presence of such a component in their source.

Although few quality Pb concentration data have been published for Hawaiian lavas, available data show that Hawaiian shield-building and transitional lavas form a steep-sloped positive array on a plot of Ce/Pb versus La/Hf that could reflect mixing between a low Ce/Pb, low La/Hf component and a high Ce/Pb, moderate La/Hf component. The low-abundance Lanai tholeiites fall at the low Ce/Pb, low La/Hf end of this array, whereas Lanai lavas containing higher REE abundances fall at the opposing end of the array, within the range observed for other

538

Hawaiian tholeiites (e.g., Kilauea, Mauna Loa). The Ce/Pb ratios of Lanai tholeiites range up to approx. 33, slightly higher than the oceanic basalt average of 25 _+ 5 suggested by Hofmann (1986); basalts from Mauna Kea are even more extreme (mean Ce/Pb = 40 + 4; Kennedy et al. 1991).

The nature of the high Ce/Pb, moderate La/Hf source component in Hawaiian lavas is problematic. Kennedy et al. (1991) observed that arc lavas have low Ce/Pb ratios that fall between primitive/oceanic crust and continental crust/sediment values (e.g., White and Patchett 1984; White and Dupre 1986) and speculated that comple- mentary arc residues could have high Ce/Pb. On the basis of those observations, Kennedy et al. (1991) suggested that the high Ce/Pb component in the source of Hawaiian lavas could consist of recycled residual subduction zone materials. However, such a component would also likely have low La/Hf. The low REE/HFSE and Th/HFSE ratios of the low-abundance Lanai tholeiites suggest that it is the low Ce/Pb Hawaiian source component that may be related to recycled subduction zone materials.

The high Ce/Pb component is unlikely to consist of subducted or delaminated continental crustal materials, because Ce/Pb ratios for average continental crust (3.~4.1; Taylor and McLennan 1985), most lower crustal xenoliths (2.3 13; Rudnick and Taylor 1986; Rudnick et al. 1986; Rudnick and Goldstein 1990), and oceanic sediments ( < 5; e.g., Ben Othman et al. 1989, McLennan et al. 1990) are even lower than primitive values. Average oceanic crust (Ce/Pb -- 14.4; Taylor and McLennan 1985) and average MORB (Ce/Pb = 25; Hofmann 1988; Sun and McDonough 1989) values are also too low for such a component. Therefore, it is unlikely this component is related to contamination of ascending Hawaiian plume melts through oceanic lithosphere and crust. Hawkesworth et al. (1991) suggest that altered MORB may have relatively high Ce/Pb ratios (approx. 42). How- ever, Mauna Kea lavas have Ce/Pb ratios as high as approx. 46 (Kennedy et al. 1991). Moreover, altered oceanic basalts cored 200 km west of the island of Hawaii have low Ce/Pb ratios (Ce/Pb < 18; Garcia unpub.). Also, isotopic data for Mauna Kea tholeiites preclude extensive interaction with MORB materials (Kennedy et al. 1991). The high Ce/Pb ratios of the Mauna Kea lavas are also too high to be related to a HIMU mantle component (Ce/Pb ~ 32; McDonough 1991). Therefore, the high Ce/Pb, moderate La/Hf component in Hawaiian lavas is not derived from these sources. Perhaps high Ce/Pb and moderate La/Hf ratios are an intrinsic feature of the Hawaiian plume source. The recycled subduction zone component could exist as blobs dispersed within this material. Positive Eu anomalies in the Hawaiian source mantle. Hofmann et al. (1984) noted that lavas with low REE abundances from the 1969-1971 Mauna Ulu eruption of Kilauea have small, positive Eu anomalies. Positive Eu anomalies are also present in some Puu Oo lavas erupted in 1983 1986 (A.W. Hofmann, personal communication). These lavas do not contain plagioclase. Although plagioclase accumulation is the most commonly invoked explanation for positive Eu anomalies (e.g., Hawkesworth et al. 1977; Woodhead 1988; Hoover 1989), clearly this is not

the case for Kilauea. Similarly, geochemical evidence ar- gues strongly against such an origin for the positive Eu anomalies in the low-abundance Lanai tholeiites. Because neither Kilauea nor Lanai tholeiites appear to have been affected by plagioclase accumulation or assimilation, these positive Eu anomalies likely are an intrinsic feature of the Hawaiian plume source.

Why are Eu anomalies absent in Lanai and Kilauea lavas containing higher REE abundances? Absolute REE abundances in Lanai tholeiites are correlated positively with La/Sm and La/Yb and correlated negatively with Eu/Eu*. Also, Eu/Eu* is correlated with highly incompatible element ratios that reflect source composition (e.g., Nb/La, La/Th, Ce/Pb; see Fig. 10). The positive Eu anomalies in the low-abundance tholeiites are accompanied by low Ce/Pb and low REE/HFSE and must, therefore, be a feature of the low Ce/Pb-low La/Hf component in the Lanai source. If the low REE/HFSE ratios in the low- abundance tholeiites are related to a subducted slab component in the Lanai source, then positive Eu and Sr anomalies may also characterize that component. It is, however, unclear what process(es) could produce a mantle component possessing positive Eu and Sr anomalies.

Temporal changes in source composition

Progressive upsection changes in incompatible trace element ratios (e.g., La/Hf, Ce/Pb, Nb/La, Sr/Nd, Zr/Sm, La/Th; see Fig. 8) are observed for lavas collected from the 2 thickest stratigraphic sections sampled on Lanai. These geochemical variations suggest that either the source of these lavas evolved geochemically with time or the processes involved in generating these lavas varied with time. Lavas with low Ce/Pb and La/Hf and high Nb/La, Eu- Eu*, and Sr/Nd appear to have erupted preferentially during the earlier stages of subaerial shield formation. If these low-abundance tholeiites represent higher degrees of partial melting than do other (younger) Lanai lavas, as suggested earlier in this paper, then these temporal geochemical changes could indicate a progressive decrease in the extent of melting in the Lanai source. These results are consistent with the findings of Kurz and Kammer (1991) showing that the source of Mauna Loa tholeiites has been changing progressively with time. We plan to sample a thicker section of Lanai lavas in more detail to further evaluate our findings.

Scale of geochemical heterogeneity

Several authors have discussed the possibility that the extent of geochemical heterogeneity in MORB is correlated inversely with the degree of partial melting in their source (Batiza 1981, 1984; Macdougall and Lugmair 1985), and a similar correlation for OIB has also been suggested (Gerlach 1990). These geochemical features have been attributed to melting of a two-component source consisting of small, enriched heterogeneities dispersed in more refractory host mantle (Sleep 1984; Batiza 1984; Macdougall and Lugmair 1985). Where the scale of heterogeneity is smaller than the scale of melting, homogeneous melts are generated. Where the scale of hetero-

539

geneity is larger than the scale of melting, heterogeneous magmas are produced. The salient consequence of this model is that variable degrees of partial melting of such a source could explain the more heterogeneous nature of OIB versus MORB and of small-volume versus large- volume hotspot volcanoes (Gerlach 1990).

The large ranges in incompatible trace element ratios in Lanai tholeiites show that source heterogeneities persist at the amounts of partial melting required to produce the substantial volumes of Hawaiian shield volcanoes. This supports the suggestion that heterogeneities in the Hawaiian plume source are significantly larger than the scale of partial melting for Hawaiian tholeiites (West et al. 1987). As discussed earlier, the geochemistry of Lanai tholeiites indicates that the source of these lavas could consist of distinct components with different solidus tem- peratures. If the Lanai source components exist as large- scale compositional heterogeneities, it implies that magmas spanning a significant fraction (if not the entire spectrum) of the compositional range of these components can potentially be produced over the duration of Hawaiian shield-building activity at each volcano, pro- vided there are sufficient variations in the degree of partial melting.

Relationship between extent of partial melting and shield volume

The ultimate cause of the large variation in the volumes of Hawaiian shield volcanoes is unknown, but it must be related to the extent of partial melting, the volume of plume material available, or a combination of the two. The geochemical characteristics of Lanai lavas, parti- cularly those of the low-abundance tholeiites, argue strongly against their having formed by smaller degrees of partial melting than that which formed tholeiites from the larger volume Hawaiian shield volcanoes. Therefore, the small volume of the Lanai shield cannot be explained by smaller degrees of partial melting of Hawaiian plume material. This suggests that differences in shield volume between Hawaiian volcanoes must be controlled primarily by differences in the volume of plume material feeding individual volcanoes, rather than by differences in the extcnt of melting of similar volumes of plume material. Thus, the flux of plume material from the Hawaiian hotspot source has varied significantly over the lifetime of the Hawaiian hotspot.

The variation in Hawaiian shield volumes could result from differences in the volume of discrete blobs of plume material feeding each volcano or from changes in thc rate plume material was supplied from the source of these lavas. Relatively large degrees of partial melting of a small-volume blob supplying Lanai could have resulted in a premature exhaustion of plume material. This could account for the cessation of volcanism at a relatively early stage in the evolution of the Lanai shield.

Conclusions

Tholeiites from the island of Lanai exemplify one isotopic extreme among Hawaiian lavas (West et al. 1987). Major

and trace element data for new samples collected from this volcano show that some Lanai lavas are unique geochemically among Hawaiian lavas. Relative to other Hawaiian tholeiites, these unique Lanai lavas contain significantly lower abundances of many incompatible elements (e.g., REE, Th) and have lower Ce/Pb, La/Yb, Th/HFSE, REE/HFSE, and alkali/HFSE ratios. These low-abundance tholeiites also possess positive Eu anomalies that are not the result of plagioclase accumulation or assimilation and probably are a feature of the Lanai source.

The geochemical characteristics of the low-abundance Lanai tholeiites indicate they may have formed by larger degrees of partial melting than did other Hawaiian tholeiites. The large range of highly incompatible trace element ratios in these lavas, along with ratio-ratio and ratio- abundance correlations, can be explained by variable degrees of melting of a geochemically heterogeneous source containing distinct components with different solidi. These source hetcrogeneities must have been significantly larger than the scale of partial melting. The low REE/HFSE Lanai source component may consist of subducted slab material in the Hawaiian plume source. The systematic temporal geochemical variations in Lanai tholeiites could have resulted from a progressive decrease in the extent of partial melting which, in turn, controlled the relative contribution of geochemically distinct source components to generated melts.

There is no positive correlation between the volumes of Hawaiian shield volcanoes and the apparent extent of partial melting. Therefore, the volumes of individual Hawaiian shields must be controlled primarily by the volume of plume material supplying each volcano. A small volume of supplied plume material, combined with relatively large degrees of partial melting, may explain why Lanai ceased eruptive activity during the shield-building stage.

Acknowledgements. We thank G.P. Russ and J. Bazan for their generous assistance and access to the LLNL ICP-MS facilities, and T.P. Hulsebosch for his invaluable kokua with XRF analyses. We thank Jo Ann Sinton for her expert thin-section work and K. Spencer for his isotope dilution analysis. Mahalo nui loa to M.D. Norman for constructive and holistic reviews of two versions of this manuscript. We acknowledge detailed, critical reviews by F.A. Frey and an anonymous reviewer that resulted in an improved paper. Work on Lanai could not have been completed without the gracious permission of Kaululaau. Research was supported in part by Na- tional Science Foundation grants EAR90-18592 to HBW and OCE90-12030 to MOG. SOEST Contribution No. 2966.

References

Andcrs E, Grevesse N (1989) Abundances of the elements: meteoritic and solar. Geochim Cosmochim Acta 53 : 197 214

Bargar KE, Jackson ED (1974) Calculated volumes of individual shield volcanoes along the Hawaiian-Emperor Chain. USGS J Res 2 : 545-550

BVSP (Basaltic Volcanism Study Project) (1981) Basaltic volcanism on the terrestrial planets. Pergamon Press Inc, New York USA

Batiza R (1981) Lithospheric age dependence of off-ridge volcano production in the North Pacific. Geophys Res Lett 8 : 853-856

Batiza R (1984) Inverse relationship between Sr isotope diversity and rate of oceanic volcanism has implications for mantle heterogeneity. Nature 309:440 441

540

Baxter AN Upton BGJ, White WM (1985) Petrology and geochemistry of Rodrigues Island, Indian Ocean. Contrib Mineral Petrol 89 : 90.101

Ben Othman D, White WM, Patchett J (1989) The geochemistry of marine sediments, island arc magma genesis, and crust-mantle recycling. Earth Planet Sci Lett 94:1-21

Bonhommet N, Beeson MH, Dalrymple GB (1977) A contribution to the geochronology and petrology of the Island of Lanai, Hawaii. GSA Bull 88:1282 1286

Boyd FR, Nixon PH (1972) Ultramafic nodules from the Thaba Putsoa Kimberlite pipe. Cam Inst Wash Ybook 71 : 362-373

Budahn JR, Schmitt RA {1985) Petrogenetic modeling of Hawaiian tholeiitic basalts: a geochemical approach. Geochim Cosmochim Acta 49 : 67-87

Campbell AC, Palmer MR, Klinkhammer GP, Bowers TS, Edmond JM, Lawrence JR, Casey JF, Thompson G, Humphris S, Rona P, Karson JA (1988) Chemistry of hot springs on the Mid-Atlantic Ridge. Nature 335:514-519

CasadevaU TJ, Dzurisin D (1987a) Intrusive rocks of Kilauea caldera. USGS Prof Pap 1350:377-394

Casadevall T J, Dzurisin D (1987b) Stratigraphy and petrology of the Uwekahuna Bluff Section, Kilauea Caldera. USGS Prof Pap 1350:351-375

Chert CY, Frey FA (1983) Origin of Hawaiian tholeiites and alkalic basalt. Nature 302:785 789

Chen CY, Frey FA (1985) Trace element and isotopic geochemistry of lavas from Haleakala Volcano, East Maul, Hawaii. Implica- tions for the origin of Hawaiian basalts. J Geophys Res 90: 8743-8768

Chen CY, Frey FA, Garcia MO, Dalrymple GB, Hart SR (1991) The tholeiite to alkalic basalt transition at Haleakala Volcano, Maul, Hawaii. Contrib Mineral Petrol 106:183 200

Clague DA, Frey FA (1982) Petrology and trace element geochemistry of the Honolulu Volcanics, Oahu: implications for the oceanic mantle below Hawaii. J Petrol 23 : 447-504

Clague DA, Weber WS, Dixon JE (1991) Picritic glasses from Hawaii. Nature 353 : 553 556

Duffy TS, Anderson DL (1989) Seismic velocities in mantle minerals and the mineralogy of the upper mantle. J Geophys Res 94 : 1895-1912

Dupuy C, Dostal J, Marcelot G, Bougault H, Joron JL, Treuil M (1982) Geochemistry of basalts from central and southern New Hebrides Arc: implication for their source rock composition. Earth Planet Sci Lett 60:207 225

Dzurisin D, Koyanagi RY, English TT (1984) Magma supply and storage at Kilauea Volcano, Hawaii, 1956-1983. J Vole Geoth Res 21 : 177-206

Eaton JP (1962) Crustal structure and volcanism in Hawaii. AGU Geophys Monograph 6:13~9

Eaton JP, Murata KJ (1960) How volcanoes grow. Science 132:925 938

Elthon D (1984) Plagioclase buoyancy in oceanic basalts: chemical effects. Geochim Cosmochim Acta 48:753 768

Falloon TJ, Green DH, Hatton CJ, Harris KL (1988) Anhydrous partial melting of a fertile and depleted peridotite from 2 to 30 kb and application to basalt petrogenesis. J Petrol 29:1257 1282

Feigenson MD (1984) Geochemistry of Kauai volcanics and a mixing model for the origin of Hawaiian alkali basalts. Contrib Mineral Petrol 87:109 119

Feigenson MD (1986) Constraints on the origin of Hawaiian lavas. J Geophys Res 91:9383-9393

Feigenson MD, Hofmann AW, Spera FJ (1983) Case studies on the origin of basalt II. The transition from tholeiitie to alkalic volcanism on Kohala Volcano, Hawaii. Contrib Mineral Petrol 84:390-405

Frey FA (1984) Rare earth element abundances in upper mantle rocks. In: Henderson P (ed) Rare earth element geochemistry. Elsevier, Amsterdam New York, pp 153 203

Frey FA, Roden MF (1987) The mantle source for the Hawaiian Islands. Constraints from the lavas and ultramafic inclusions. In: Menzies M, Hawkesworth C (eds) Mantle metasomatism. Aca- demic Press, New York London, pp 423-463

Frey FA, Wise WS, Garcia MO, West H, Kwon S-T, Kennedy A (1990a) Evolution of Mauna Kea Volcano, Hawaii: petrologic and geochemical constraints on postshield volcanism. J Geophys Res 95 : 1271-1300

Frey FA, Garcia MO, Chen CY, Wise WS, Roden M, Kennedy A, West H (1990b) Geochemical constraints on the evolution of Hawaiian volcanoes. In: Intraplate volcanism: the Reunion Hot Spot. Abstr Vol p 77

Fujimaki H, Tatsumoto M, Aoki KI (1984) Partition coefficients of Hf, Zr, and REE Between phenocrysts and groundmass. Proc 14 th Lunar Planet Sci Conf: 66~672

Gareia MO, Hulsebosch TP (1990) Neutral buoyancy and the origin of submarine picrites at Kilauea Volcano, Hawaii. In: Intraplate volcanism: the Reunion Hot Spot, Abstract Vol. p 79

Gerlach DC (1990) Eruption rates and isotopic systematics of ocean islands: further evidence for small-scale heterogeneity in the upper mantle. Tectonophys 172 : 273 289

Gill J (1981) Orogenic andesites and plate tectonics, Springer, Heidelberg New York London Tokyo, 390 p

Gladney ES, Roelandts I (1988) 1987 Compilation of elemental concentration data for USGS BHVO-1, MAG-1, QLO-1, RGM- 1, SCo-1, SGR-1 and STM-1. Geostandards Newsletter 12:253 362

Govindaraju K (1989) 1989 Compilation of working values and sample description for 272 geostandards. Geostandards News- letter 13:1-113

Green DH (1973) Experimental melting studies on a model upper mantle composition at high pressure under water-saturated and water-undersaturated conditions. Earth Planet Sci Lett 19 : 37-53

Green DH, Ringwood AE (1967) The genesis of basaltic magmas. Contrib Mineral Petrol 15 : 103-190

Hart SR (1969) K, Rb, Cs Contents and K/Rb, K/Cs Ratios of fresh and altered submarine basalts. Earth Planet Sci Lett 6:295 303

Hart SR, Staudigel H (1982) The control of alkalies and uranium in seawater by ocean crust alteration. Earth Planet Sci Lett 58:202 212

Hawkesworth C J, O'Nions RK, Pankhurst R J, Hamilton P J, Evensen NM (1977) A geochemical study of island-arc tholeiites from the Scotia Sea. Earth Planet Sci Lett 36:353-362

Hawkesworth CJ, Hergt JM, Ellam RM, McDermott F (1991) Element fluxes associated with subduction related magmatism. Phil Trans R Soc London A-335 : 393 405

Hegner E, Unruh D, Tatsumoto M (1986) Nd-Sr-Pb isotope constraints on the sources of West Maul Volcano, Hawaii. Nature 319:478-480

Higuchi H, Nagasawa H (1969) Partition of trace elements between rock forming minerals and the host volcanic rocks. Earth Planet Sci Lett 7:281~87

Ho RA, Garcia MO (1988) Origin of differentiated lavas at Kilauea Volcano, Hawaii: implications from the 1955 eruption. Bull Volc 50 : 35-46

Hofmann AW (1986) Nb in Hawaiian magmas: constraints on source composition and evolution. Chem Geol 57 : 17-30

Hofmann AW (1988) Chemical differentiation of the Earth: the relationship between mantle, continental crust, and oceanic crust. Earth Planet Sci Lett 90:297 314

Hofmann AW, White WM (1983) Ba, Rb and Cs in the Earth's mantle. Z Naturforsch 38a:25(~266

Hofmann AW, Feigenson MD, Raczek I (1984) Case studies on the origin of basalt: III petrogenesis of the Mauna Ulu eruption, Kilauea, 1969 1971. Contrib Mineral Petrol 88:24-35

Hofmann AW, Jochum KP, Seufert M, White WM (1986) Nb and Pb in oceanic basalts: new constraints on mantle evolution. Earth Planet Sci Lett 79:33-45

Hoover JD (1989) The chilled marginal gabbro and other contact rocks of the Skaergaard Intrusion. J Petrol 30:44'1-476

Irvine TN (1977) Definition of primitive liquid compositions for basic magmas. Carn Inst Wash Ybook 76:454-461

Jackson ED, Wright TL (1970) Xenoliths in the Honolulu Volcanic Series, Hawaii. J Petrol 11:405 430

Jaques AL, Green DH (1980) Anhydrous melting of peridotite at

541

0-15 kb pressure and the genesis of tholeiitic basalts. Contrib Mineral Petrol 73:287-310

Kennedy AK, Kwon ST, Frey FA, West HB (1991) The isotopic composition of postshield lavas from Mauna Kea Volcano, Hawaii. Earth Planet Sci Lett 103:339 353

Klein FW, Koyanagi RY, Nakata JS, Tanigawa WR (1987) The seismicity of Kilauea's magma system. USGS Prof Pap 1350:1019-1185

Kurz MD, Kammer DP (1991) Isotopic evolution of Mauna Loa Volcano. Earth Planet Sci Lett 103:257 269

Kurz MD, Jenkins WJ, Hart SR, Clague D (1983) Helium isotopic variations in volcanic rocks from Loihi Seamount and the Island of Hawaii. Earth Planet Sci Lett 66 : 388-406

Lanphere MA, Frey FA (1987) Geochemical evolution of Kohala Volcano, Hawaii. Contrib Mineral Petrol 95:100 113

Leeds AR (1975) Lithospheric thickness in the western Pacific. Phys Earth Planet lnt 11:61-64

Leeman WP, Budahn JR, Gerlach DC, Smith DR, Powcll BN (1980) Origin of Hawaiian tholeiites: trace element constraints. Am J Sci 280-A : 794 819

Leeman WP, Gerlach DC, Garcia MO, West HB (1992) Geochemi- cal variations in lavas from Kahoolawe Volcano, Hawaii. Con- trib Mineral Petrol (in press)

LeRoex AP, Erlank AJ (1982) Quantitative evaluation of fractional crystallization in Bouvet Island lavas. J Vole Geoth Res 13:309 338

Lipman PW, Banks NG, Rhodes, JM 1985, Degassing-induced crystallization of basaltic magma and effects on lava rheology: Nature 317, p. 604-607

Liu M, Chase CG (1989) Evolution of midplate hotspot swells: numerical solutions. J Geophys Res 94:5571 5584

Lockwood JP, Dvorak JJ, English TT, Koyanagi RY, Okamura AT, Summers ML, Tanigawa WR (1987) Mauna Loa 1974-1984: A decade of intrusive and extrusive activity. USGS Prof Pap 1350:537-570

Ludden JN, Thompson G (1979) An evaluation of the behavior of the rare earth elements during the weathering of sea-floor basalt. Earth Planet Sci Lett 43:85-92

Maaloe S (1979) Compositional range of primary tholeiitic magmas evaluated from major-element trends. Lithos 12:59-72

Maaloe S, Jakobsson SP (1980) The PT phase relations of a primary oceanite from the Reykjanes Peninsula, Iceland. Lithos 13 : 237 246

Maaloe S, Hansen B (1982) Olivine phenocrysts of Hawaiian olivine tholeiite and oceanite. Contrib Mineral Petrol 81:203-211

Macdonald GA (1968) Composition and origin of Hawaiian lavas. GSA Mere 116:477-522

Macdonald GA, Katsura T (1961) Variations in the lava of the 1959 eruption in Kilauea Iki. Pac Sci 15:358 369

Macdonald GA, Katsura T (1964) Chemical composition of Ha- waiian lavas. J Petrol 5:82-133

Macdougall JD, Finkel RC, Carlson J (1979) Isotopic evidence for uranium exchange during low-temperature alteration of oceanic basalt. Earth Planet Sci Lett 42 : 27-34

Macdougall JD, Lugmair GW (1985) Extreme isotopic homogeneity among basalts from the southern East Pacific Rise: mantle or mixing effect? Nature 313:209 211

McDonough WF (1991) Partial melting of subducted oceanic crust and isolation of its residual eclogitic lithology. Phil Trans R Soc Lond 335-A : 407-418

McLennan SM, Taylor SR, McCulloch MT, Maynard JB (1990) Geochemical and Nd-Sr isotopic composition of deep-sea turbi- dites: crustal evolution and plate tectonic associations. Geochim Cosmochim Acta 54:2015-2050

Michard A, Albarede F (1986) The REE content of some hydrother- real fluids. Chem Geol 55:51-60

Michard A, Albarede F, Michard G, Minster JF, Charlou JL (1983) Rare-earth elements and uranium in high-temperature solutions from east Pacific Rise hydrothermal vent field (13~ Nature 303 : 795-797

Moore JG (1987) Subsidence of the Hawaiian ridge. USGS Prof Pap 1350:85-100

Moore RB, Helz RT, Dzurisin D, Eaton GP, Koyanagi RY, Lipman PW, Lockwood JP, Puniwai GS (1980) The 1977 Eruption of Kilauea Volcano, Hawaii. J Volc Geoth Res 7:189-210

Morse SA (1991) Basaltic magma from the crust is not a free option. EOS 72 : 161

Morse SA, Nolan KM (1985) Kiglapait geochemistry VII: yttrium and the rare earth elements. Geochim Cosmochim Acta 49 : 1621 1644

Naughton JJ, Macdonald GA, Greenberg VA (1980) Some addi- tional potassium-argon ages of Hawaiian rocks: the Maul Vol- canic Complex of Molokai, Maui, Lanai and Kahoolawe. J Volc Geoth Res 7 : 339-355

Newsom HE, White WM, Jochum KP, Hofmann AW (1986) Siderophile and chalcophile element abundances in oceanic basalts, Pb isotope evolution and growth of the Earth's core. Earth Planet Sci Lett 80 : 299-3 J 3

Nicholls J, Stout MZ (1988) Picritic melts in Kilauea evidence from the 1967-1968 Halemaumau and Hiiaka eruptions. J Petrol 29:1031 1057

Norrish K, Chappell BW (1977) X-Ray fluorescence spectrometry. In: Zussman J (ed) Physical methods in determinative mineralogy. Academic Press, New York London, pp 201-272

O'Hara M J, Richardson SW, Wilson G (1971) Garnet-peridotite stability and occurrence in crust and mantle. Contrib Mineral Petrol 32:48 68

Paster TP, Schauwecker DS, Haskin LA (1974) The behavior of some trace elements during solidification of the SkaerKaard layered series. Geochim Cosmochim Acta 38:1549-1577

Philpotts JA, Scbnetzler CC (1970) Phenocryst-matrix partition coefficients for K, Rb, Sr and Ba, with applications to anorthosite and basalt genesis. Geochim Cosmochim Acta 34:307 322

Phinney WC, Morrison DA (1990) Partition coefficients for calcic plagioclase: implications for Archean anorthosites. Geochim Cosmochim Acta 54 : 1639 1654

Prinzhofer A, Allegre CJ (1985) Residual peridotites and the mechanisms of partial melting. Earth Planet Sci Lett 74:25l~65

Rhodes JM (1968) Geochemistry of the 1984 Mauna Loa eruption: implications for magma storage and supply. J Geophys Res 93 : 4453-4466

Roden MF, Frey FA, Clague DA (1984) Geochemistry of tholeiitic and alkalic lavas from the Koolau Range, Oahu, Hawaii. Implications for Hawaiian volcanism. Earth Planet Sci Lett 69:141 158

Rudnick RL, Taylor SR (1986) The composition and petrogenesis of the lower crust: a xenolith study. J Geophys Res 92 : 13981-14005

Rudnick RL, Goldstein SL (1990) The Pb isotopic compositions of lower crustal xenoliths and the evolution of lower crustal Pb. Earth Planet Sci Lett 98 : 192-207

Rudnick RL, McDonough WF, McCulloch MT, Taylor SR (1986) Lower crustal xenoliths from Queensland Australia: evidence for deep crustal assimilation and fractionation of continental basalts. Geochim Cosmochim Acta 50 : 1099 1115

Ryan MP, Koyanagi RY, Fiske RS (1981) Modeling the three- dimensional structure of macroscopic magma transport systems: application to Kilauea Volcano, Hawaii. J Geophys Res 86:7111 7129

Saunders AD, Norry M J, Tarney J (1980) Transverse geochemical variations across the Antarctic Peninsula: implications for the genesis of calc-alkaline magmas. Earth Planet Sci 46: 344-360

Saunders AD, Norry MJ, Tarney J (1991) Fluid influence on the trace element compositions of subduction zone magmas. Phil Trans R Soe Lond 335-A:377-392

Schnetzler CC, Philpotts JA (1970) Partition coefficients of rare- earth elements between igneous matrix material and rock-forming mineral phenocrysts-II. Geocheim Cosmochim Acta 34 : 331-340

Shaw DM (1970) Trace element fraetionation during anatexis. Geo- chim Cosmochim Acta 34:237 243

Shimizu N (1975) Rare earth elements in garnets and clinopyroxenes from garnet lherzolite nodules in kimberlites. Earth Planet Sci Lett 25 : 26-32

Sleep NH (1984) Tapping of magmas from ubiquitous mantle

542

heterogeneities: an alternative to mantle plumes? J Geophys Res 89 : 10029-10041

Staudigel H, Zindler A, Hart SR, Leslie T, Chen C-Y, Clague DA (1984) The isotope systematics of a juvenile intraplate volcano: Pb, Nd, and Sr Isotope ratios of basalts from Loihi Seamount, Hawaii. Earth Planet Sci Lett 69 : 1.3-29

Stearns HT (1940) Geology and ground-water resources of Lanai and Kahoolawe Hawaii. Hawaii Div Hydrogr Bull 6, 177 pp

Stille P, Unruh DM, Tatsumoto M (1983) Pb, St, Nd And Hf isotopic evidence of multiple sources for Oahu, Hawaii basalts. Nature 304 : 25-29

Stille P, Unruh DM, Tatsumoto M (1986) Pb, Sr, Nd, and Hf isotopic constraints on the origin of Hawaiian basalts and evidence for a unique mantle source. Geochim Cosmochim Acta 50:2303 2319

Sun SS, McDonough WF (1989) Chemical and isotopic systematics of oceanic basalts: implications for mantle composition and Processes. In: Saunders AD, Norry MJ (eds) Magmatism in the ocean basins. Blackwell Scientific Publishers, Oxford London, pp. 313 345

Tatsumoto M (1978) Isotopic composition of lead in oceanic basalt and its implication to mantle evolution. Earth Planet Sci Lett 38 : 62-87

Tatsumoto M, Hegner E, Unruh DM (1987) Origin of the west Maui volcanic rocks inferred from Pb, Sr, and Nd isotopes and a multicomponent model for oceanic basalt. USGS Prof Pap 1350:723-744

Taylor SR, McLennon SM (1985) The continental crust: its composition and evolution. Blackwell Scientific Publishers, Oxford London

Taylor WR, Green DH (1988) Measurement of reduced peridotite- C-O-H solidus and implications for redox melting of the mantle. Nature 332 : 349-352

Thompson RN, Morrison MA, Hendry DL, Parry SJ (1984) An assessment of the relative roles of crust and mantle in magma genesis: an elemental approach. Phil Trans R Soc London A- 310:549-590

Tilling RI, Wright TL, Millard HT (1987) Trace-element chemistry of Kilauea and Mauna Loa in space and time: a reconnaissance. USGS Prof Pap 1350 : 641 689

Washington HS (1923) Petrology of the Hawaiian Islands: I. Kohala and Mauna Kea. Am J Sci 5:465-502

Watts AB (1978) An analysis of isostasy in the world's oceans 1. Hawaiian-Emperor Seamount Chain. J Geophys Res 83 : 5989-6004

Weaver BL (1991) The origin of ocean island basalt end-member compositions: trace element and isotopic constraints. Earth Pla- net Sci Lett 104; 381-396

West HB, Leeman WP (1987) Isotopic evolution of lavas from Haleakala Crater, Hawaii. Earth Planet Sci Lett 84: 2l 1-225

West HB, Gerlach DC, Leeman WP, Garcia MO (1987) Isotopic constraints on the origin of Hawaiian magmas from the Maul Volcanic Complex, Hawaii. Nature 330:216-220

White WM (1985) Sources of oceanic basalts: radiogenic isotopic evidence. Geology 13 : 115-118

White WM, Patchett J (1984) Hf-Nd-Sr isotopes and incompatible element abundances in island arcs: implications for magma origins and crust-mantle evolution. Earth Planet Sci Lett 67:167 185

White WM, Dupre B (1986) Sediment subduction and magma genesis in the Lesser Antilles: isotopic and trace element constraints. J Geophys Res 91:5927-5941

Wilkinson JFG (1985) Undepleted mantle composition beneath Hawaii. Earth Planet Sci Lett 75 : 129-138

Wilkinson JFG, Hensel HD (1988) The petrology of some picrites from Mauna Loa and Kilauea volcanoes, Hawaii. Contrib Min- eral Petrol 98 : 326-345

Woodhead JD (1988) The origin of geochemical variations in Mar- iana lavas: a general model for petrogenesis in intra-oceanic island arcs? J Petrol 29:805-830

Wright TL (1971) Chemistry of Kilauea and Mauna Loa in space and time. USGS Prof Pap 735 : 1 40

Wright TL (1984) Origin of Hawaiian tholeiite: a metasomatic model. J Geophys Res 89:3233-3252

Wright TL, Fiske RS (1971) Origin of the differentiated and hybrid lavas of Kilauea Volcano, Hawaii. J Petrol 12 : 1-65

Wright TL, Swanson DA, Duflield WA (1975) Chemical compositions of Kilauea East Rift lava, 1968-1971. J Petrol 16:110-133

Zindler A, Hart SR (1986) Chemical geodynamics. Ann Rev Earth Planet Sci 14: 493-571

Editorial responsibility. J. Patchett

Documents

Geochemistry of tholeiites from Lanai, Hawaii...bedded (0.5-1 m thick), highly vesicular pahoehoe flows. We also collected a sample from Poaiwa, located on the northeast coast of Lanai,